CN106156727A - The recognition methods of a kind of biological characteristic and terminal - Google Patents

Abstract

The embodiment of the present invention discloses a biological feature recognition method and a terminal, which are used to accelerate the speed of identifying the biological feature by the terminal. The method in the embodiment of the present invention includes: the terminal determines the target equal error rate and the target false recognition rate; the terminal reduces the dimensionality of the first original feature of the first sample according to the target equal error rate to obtain the first feature, and according to the target false recognition rate the second An original feature is dimensionally reduced to obtain the second feature, the first sample is a sample stored or collected in advance by the terminal; the terminal stores the first feature and the second feature; the terminal obtains the biological image of the object to be identified to generate the second sample and extracts the second feature The second original feature of two samples; the terminal reduces the dimension of the second original feature according to the target equal error rate to obtain the third feature, and reduces the dimension of the second original feature according to the target false recognition rate to obtain the fourth feature; the terminal combines the third feature with The first feature is compared and recognized, and the fourth feature is compared and recognized with the second feature to obtain a recognition result.

Description

A biometric identification method and terminal

technical field

The invention relates to the field of terminals, in particular to a biometric identification method and a terminal.

Background technique

Biometric technology is a technology that uses human biological characteristics for identity authentication. Biometric characteristics are unique (different from others), physiological characteristics or behaviors that can be measured or automatically identified and verified, and are divided into physiological characteristics and behavioral characteristics. The biometric system samples the characteristic characteristics, extracts its unique characteristics and performs identity authentication. As a means of identification, biometric technology has unique advantages, and has gradually become a research hotspot in the world in recent years. At present, the methods that are relatively mature and used on a large scale mainly include fingerprints, irises, faces, ears, palm prints, and hand stamp veins. Biometric identification technology usually follows several steps of scanning, digital processing, analysis, feature extraction, storage, matching and classification.

At present, in order to obtain higher reliability in biometric recognition, usually high-dimensional features are extracted as much as possible during the feature extraction stage. For example, after the face is divided into 9x9 blocks to extract LBP features and calculate the histogram, the obtained features are 9x9x256 original Local binary pattern (full name in English: Local Binary Patterns, referred to as LBP) features, if you use a unified LBP (full name in English: Unified LBP), its features are 9x9x58; the features obtained by the feature extraction algorithm based on key point processing blocks will also Many, for example, 7 key points, the Unified LBP histogram features obtained by the 4x4 block method at each key point are 7x4x4x58=6496.

However, the high-dimensional features are large in size and have many features, which leads to a complicated calculation process in the comparison and identification of the terminal, which in turn reduces the speed of the terminal's identification of biometric features.

Contents of the invention

Embodiments of the present invention provide a biometric feature identification method and a terminal, which are used to speed up the identification of biometric features by the terminal.

In a first aspect, an embodiment of the present invention provides a biometric identification method, including:

The terminal determines the target equal error rate and the target false recognition rate; the terminal reduces the dimensionality of the first original feature of the first sample according to the target equal error rate to obtain the first feature, and reduces the first original feature according to the target false recognition rate. The first sample is a sample stored or collected in advance by the terminal; the terminal stores the first feature and the second feature; the terminal obtains the biological image of the object to be identified to generate a second sample and extracting the second original feature of the second sample; the terminal reduces the dimension of the second original feature according to the target equal error rate to obtain a third feature, and reduces the dimension of the second original feature according to the target false recognition rate to obtain a fourth feature feature; the terminal compares and recognizes the third feature with the first feature and compares and recognizes the fourth feature with the second feature to obtain a recognition result.

In practical applications, the terminal reduces the dimension of the first original feature of the first sample to obtain the first feature according to the target equal error rate after the terminal reduces the dimension of the first original feature to obtain the second feature according to the target false recognition rate , the terminal then reduces the dimensionality of the second feature according to the target equal error rate to obtain the first feature, and the specific method is not limited here.

In a possible implementation manner, the target equal error rate is within the first value range so that the first feature does not exceed the minimum equal error rate of the second value range, and the first value range is the target equal error rate value range, the second value range is the value range of the first feature;

The target false recognition rate is within the third value range so that the second feature does not exceed the minimum false recognition rate of the fourth value range, the third value range is the value range of the target false recognition rate, and the fourth The value range is the value range of the second characteristic.

In another possible implementation manner, the terminal comparing and identifying the third feature with the first feature and comparing and identifying the fourth feature with the second feature to obtain a recognition result includes: the terminal judging that the third feature Whether the similarity between the feature and the first feature is less than a first threshold;

If the similarity between the third feature and the first feature is not less than the first threshold, the terminal judges whether the fourth feature and the second feature are less than a second threshold;

If the similarity between the fourth feature and the second feature is not less than a second threshold, the terminal determines that the second sample is the same as the first sample.

In another possible implementation manner, after the terminal determines whether the similarity between the third feature and the first feature is less than a first threshold, the method further includes: if the similarity between the third feature and the first feature is less than the first threshold, the terminal determines that the second sample is different from the first sample.

In another possible implementation manner, after the terminal determines whether the fourth feature and the second feature are smaller than a second threshold, the method further includes: if the similarity between the fourth feature and the second feature is smaller than the second threshold, Then the terminal judges that the second sample is different from the first sample.

In another possible implementation manner, the terminal storing the first feature and the second feature includes:

The terminal quantizes the first feature according to a relational expression to obtain a first quantized feature, and quantizes the second feature according to the relational expression to obtain a second quantized feature;

The terminal stores the first quantized feature and the second quantized feature;

The terminal compares and recognizes the third feature with the first feature and compares and recognizes the fourth feature with the second feature to obtain a recognition result including:

The terminal quantizes the third feature according to the relational expression to obtain a third quantized feature, and quantizes the fourth feature according to the relational expression to obtain a fourth quantized feature;

The terminal compares and identifies the third quantitative feature with the first quantitative feature and compares and identifies the fourth quantitative feature with the second quantitative feature to obtain a recognition result.

In practical applications, the terminal may also store a personal identification code (English full name: Personal Identification Number, PIN for short) corresponding to the first feature and the second feature.

In another possible implementation, the relationship is:

$\mathrm{<span\; class="notranslate">\; f</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; (</span>\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; V</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; N</span>}}{{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$

Wherein, V is the feature value after dimensionality reduction of the sample, V _min is the minimum feature value after dimensionality reduction of the sample, V _max is the maximum feature value after dimensionality reduction of the sample, and the value of N is 255 or 65535.

In a second aspect, an embodiment of the present invention provides a terminal, including:

A determination module is used to determine the target equal error rate and the target false recognition rate;

The first dimension reduction module is used to reduce the dimensionality of the first original feature of the first sample to obtain the first feature according to the target equal error rate determined by the determination module, and to obtain the first feature according to the target false recognition rate determined by the determination module. The first original feature is dimensionally reduced to obtain the second feature, and the first sample is a sample stored or collected in advance by the terminal;

a storage module, configured to store the first feature and the second feature obtained by reducing the dimensionality of the dimensionality reduction module;

An acquisition module, configured to acquire a biological image of an object to be identified to generate a second sample and extract a second original feature of the second sample;

The second dimension reduction module is used to reduce the dimensionality of the second original feature acquired by the acquisition module to obtain a third feature according to the target equal error rate determined by the determination module, and to obtain a third feature according to the target false recognition rate determined by the determination module. The second original feature acquired by the acquisition module is dimensionally reduced to obtain a fourth feature;

An identification module, configured to compare and identify the third feature obtained by reducing the dimensionality of the second dimensionality reduction module with the first feature obtained by reducing the dimensionality of the first dimensionality reduction module, and obtain the dimensionality reduction obtained by the second dimensionality reduction module The fourth feature is compared and identified with the second feature obtained by dimensionality reduction of the first dimensionality reduction module to obtain a recognition result.

In a possible implementation manner, the target equal error rate is within the first value range so that the first feature does not exceed the minimum equal error rate of the second value range, and the first value range is the target equal error rate value range, the second value range is the value range of the first feature;

The target false recognition rate is within the third value range so that the second feature does not exceed the minimum false recognition rate of the fourth value range, the third value range is the value range of the target false recognition rate, and the fourth The value range is the value range of the second characteristic.

In another possible implementation manner, the identification module includes:

a first judging unit, configured to judge whether the similarity between the third feature and the first feature is smaller than a first threshold;

A second judging unit, configured to judge whether the fourth feature and the second feature are smaller than a second threshold if the first judging unit judges that the similarity between the third feature and the first feature is not less than the first threshold;

The first identification unit is configured to determine that the second sample is the same as the first sample if the second judgment judges that the similarity between the fourth feature and the second feature is not less than a second threshold.

In another possible implementation manner, the identification module further includes:

The second identifying unit is configured to determine that the second sample is different from the first sample if the first judging unit judges that the similarity between the third feature and the first feature is smaller than the first threshold.

In another possible implementation manner, the identification module further includes:

The third identifying unit is configured to determine that the second sample is different from the first sample if the second judging unit judges that the similarity between the fourth feature and the second feature is less than a second threshold.

In another possible implementation manner, the storage module includes:

The first quantization unit is configured to quantize the first feature according to the relational expression to obtain a first quantization feature, and quantize the second feature according to the relational expression to obtain a second quantization feature;

a storage unit, configured to store the first quantization feature and the second quantization feature quantized by the quantization unit;

The identification module includes:

The second quantization unit is configured to quantize the third feature according to the relational expression to obtain a third quantization feature, and quantize the fourth feature according to the relational expression to obtain a fourth quantization feature;

The fourth identification unit is configured to compare and identify the third quantitative feature with the first quantitative feature and compare and identify the fourth quantitative feature with the second quantitative feature to obtain a recognition result.

In another possible implementation, the relationship is:

$\mathrm{<span\; class="notranslate">\; f</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; (</span>\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; V</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; N</span>}}{{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$

Wherein, V is the feature value after dimensionality reduction of the sample, V _min is the minimum feature value after dimensionality reduction of the sample, V _max is the maximum feature value after dimensionality reduction of the sample, and the value of N is 255 or 65535.

In a third aspect, an embodiment of the present invention provides a terminal, including:

transceivers, processors, memory and buses;

The transceiver, the processor and the memory are connected through the bus;

The processor has the following functions: determine the target equal error rate and the target false recognition rate; reduce the dimensionality of the first original feature of the first sample to obtain the first feature according to the target equal error rate determined by the determination module, and obtain the first feature according to the determined The target false recognition rate determined by the module reduces the dimensionality of the first original feature to obtain a second feature, and the first sample is a sample stored or collected in advance by the terminal;

The memory has the following functions: storing the first feature and the second feature obtained by reducing the dimensionality of the dimensionality reduction module;

The transceiver has the following functions: acquiring a biological image of an object to be identified to generate a second sample and extracting a second original feature of the second sample;

The processor has the following functions: according to the target equal error rate determined by the determination module, the second original feature obtained by the acquisition module is dimensionally reduced to obtain a third feature, and according to the target false recognition rate determined by the determination module, the The dimension reduction of the second original feature obtained by the acquisition module obtains the fourth feature; the third feature obtained by the dimension reduction of the second dimension reduction module and the first feature obtained by the dimension reduction of the first dimension reduction module are compared and identified and comparing and identifying the fourth feature obtained by dimensionality reduction of the second dimensionality reduction module with the second feature obtained by dimensionality reduction of the first dimensionality reduction module to obtain a recognition result.

It can be seen from the above technical solutions that the embodiments of the present invention have the following advantages: the terminal respectively performs dimensionality reduction on the first original feature of the first sample according to the target equal error rate and the target false recognition rate to obtain the first feature and the second feature and storage; when the terminal obtains the biological image of the object to be identified to obtain the second sample and extracts the second original feature of the second sample, the second original feature is also dimensionally reduced according to the target equal error rate and target false recognition rate to obtain The third feature and the fourth feature; when the second sample is compared and identified with the first sample, the third feature is compared with the first feature, and the fourth feature is compared with the second feature, due to the reduction of the first feature The biological characteristics of the first sample and the second sample, thereby reducing the complexity of the calculation process of the terminal in the process of comparing and identifying, thereby speeding up the speed of terminal recognition of biological characteristics.

Description of drawings

FIG. 1 is a schematic diagram of an embodiment of a biometric identification method in an embodiment of the present invention;

FIG. 2 is a schematic diagram of an embodiment of a terminal in an embodiment of the present invention;

Fig. 3 is a schematic diagram of another embodiment of a terminal in the embodiment of the present invention.

detailed description

Embodiments of the present invention provide a biometric feature identification method and a terminal, which are used to speed up the identification of biometric features by the terminal.

In order to enable those skilled in the art to better understand the solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only It is an embodiment of a part of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

The following are detailed descriptions through specific embodiments. Please refer to FIG. 1. The embodiment of the present invention provides a biometric identification method, including:

101. The terminal determines a target equal error rate and a target false acceptance rate.

The user pre-sets the value range of the high-dimensional original feature after dimensionality reduction, the value range of the equal error rate and the value range of the false recognition rate; The value of the feature after the error rate dimensionality reduction does not exceed the minimum equal error rate of the pre-selected value range. The minimum false recognition rate whose value does not exceed the pre-selected value range is the target false recognition rate.

In this embodiment, taking the original feature dimension of human face as 6496 as an example, if the value range of the feature after dimension reduction according to the equal error rate is [10,20], the value of the feature after dimension reduction according to the false recognition rate The range is [200,400], and the value range of the equal error rate is [0.03,0.05] and the value range of the false recognition rate is [0.000001,0.000003]. In practical applications, if the equal error rate is 0.05, The terminal can reduce the dimension of the original feature to 10 dimensions. When the equal error rate is 0.03, the terminal can reduce the dimension of the original feature to 20 dimensions. Then the terminal chooses 0.03 as the target equal error rate. When the false rate value is 0.000001, the terminal can reduce the original feature dimension to 400 dimensions. When the false recognition rate value is 0.000003, the terminal can reduce the original feature dimension to 200 dimensions. Then the terminal chooses 0.000001 as the target equal Error rate.

102. The terminal reduces the dimensionality of the first original feature of the first sample according to the target equal error rate to obtain the first feature, and reduces the dimensionality of the first original feature according to the target false recognition rate to obtain the second feature.

The terminal obtains the first sample that needs to be stored or collected in advance and extracts the first original feature of the first sample. The terminal performs dimensionality reduction on the first original feature according to the determined target equal error rate to obtain the first feature, and performs dimensionality reduction on the first original feature according to the determined target false recognition rate to obtain the second feature.

In practical applications, when the terminal extracts the first original feature of the face, various methods can be used. For example, after the face is divided into 9x9 blocks to extract LBP features and calculate the histogram, the obtained features are 9x9x256 original LBP features. If you use In the case of Unified LBP, its features are 9x9x58; the feature extraction algorithm based on the key point block will also get a lot of features, for example, 7 key points, the Unified LBP histogram feature obtained by the 4x4 block method at each key point The number is 7x4x4x58=6496, and the extraction method of the original feature of the sample is not limited here. In this embodiment, the original feature of the face is 6496 as an example. According to the target equal error rate and target false recognition rate selected in step 101, the first feature obtained by reducing the dimension of the face is 20 dimensions, and the second feature is 400 dimensions. .

103. The terminal stores the first feature and the second feature.

The terminal stores the dimensionally reduced first feature and the second feature.

In practical applications, the terminal can also determine the PIN corresponding to the first feature and the second feature, and store the PIN and the first feature and the second feature at the same time, so that when comparing The first feature and the second feature are read directly.

In practical applications, if the storage space of the terminal is not large, the terminal can also use the relational expression: Quantify the first feature and the second feature to obtain the first quantized feature and the second quantized feature. For example, the first feature is 20 dimensions, where the value of one dimension is 0.5, the value of the other dimension is 2, and the value of the largest dimension is 3, the value of the smallest dimension is 0.2, and the value of N is 255, then the quantization values of the features obtained by quantization according to the relational expression are as follows: the quantization value of 0.5 is 27, the quantization value of 2 is 163, and the quantization value of 3 is 255, and the quantization value of 0.2 is 0, that is, the first feature can be represented by an unsigned 8-bit integer from 0 to 255, and these unsigned 8-bit integers can be represented by one byte, then store the first feature For a quantized feature, the first quantized feature only occupies 20 bytes, which greatly reduces the occupied storage space. At the same time, in practical applications, 8-bit signed integers from -128 to 127, 16-bit unsigned integers from 0 to 65535, or 16-bit signed integers from -32768 to 32767 can also be used to represent features during quantization. The specific method is not described here. Do limited.

104. The terminal acquires the biological image of the object to be identified to generate a second sample and extracts a second original feature of the second sample.

During the use of the terminal, the biological image of the object to be recognized, such as a human face, is obtained, and after the biological image is obtained, it is used as a second sample to be identified, and the second original feature of the biological image is extracted.

In practical applications, the biological image of the object to be identified acquired by the terminal can be fingerprints, iris, face, ear, palm print, hand stamp veins, etc. At the same time, the way the terminal acquires the biological image of the object to be identified can be through the camera, infrared scanning Etc., the method for the terminal to extract the second original feature may also be to extract the LBP feature and calculate the histogram or obtain the feature based on the feature extraction algorithm of the key point processing block, which is not limited here.

105. The terminal reduces the dimensionality of the second original feature according to the target equal error rate to obtain a third feature, and reduces the dimensionality of the second original feature according to the target false recognition rate to obtain a fourth feature.

The terminal performs dimensionality reduction on the second original feature according to the determined target equal error rate to obtain a third feature, and performs dimensionality reduction on the second original feature according to the determined target false recognition rate to obtain a fourth feature.

In practical applications, if the terminal quantifies the first feature and the second feature of the pre-stored or collected first sample and stores them after the storage space is small, the terminal also needs to quantify the first feature and the second feature of the second sample The third feature and the fourth feature are quantized in the same manner to obtain quantized features.

106. The terminal compares and recognizes the third feature with the first feature, and compares and recognizes the fourth feature with the second feature to obtain a recognition result.

The terminal sets a first threshold of similarity between the third feature and the first feature, and sets a second threshold of similarity between the fourth feature and the second feature, and then the terminal judges that the third feature is similar to the second feature Whether the similarity of the first feature is less than the first threshold; if the similarity between the third feature and the first feature is not less than the first threshold, then the terminal judges whether the fourth feature and the second feature are smaller than the second A threshold; if the similarity between the fourth feature and the second feature is not less than a second threshold, the terminal determines that the second sample is the same as the first sample. If the similarity between the third feature and the first feature is less than the first threshold, the terminal judges that the second sample is different from the first sample, and at the same time, it is not necessary to compare and identify the fourth feature and the second feature. If the similarity between the third feature and the first feature is not less than the first threshold and the similarity between the fourth feature and the second feature is less than the second threshold, the terminal also judges that the The second sample is different from the first sample.

For example, the first threshold is 0.8, the second threshold is 0.9, the similarity between the third feature and the first feature is 0.85, and when the similarity between the fourth feature and the second feature is 0.9, the terminal judges that the second sample is the same as the first This is the same person, if the similarity between the third feature and the first feature is 0.79, then the terminal judges that the second sample and the first sample are not the same person or the similarity between the third feature and the first feature is 0.85, then If the similarity between the fourth feature and the second feature is 0.89, the terminal also determines that the second sample is not the same person as the first sample.

In practical applications, if the terminal quantifies the first feature and the second feature of the pre-stored first sample and stores them after the storage space is small, then the terminal needs to use the The third quantitative feature and the first quantitative feature, and the fourth quantitative feature and the second quantitative feature are identified using the same comparison and identification method as described above.

For ease of understanding, the biometric identification method in this embodiment is described in detail below using a practical application scenario. In this embodiment, terminals include but are not limited to mobile phones, punch cards, etc., and punch cards are taken as examples of terminals.

If the punch card machine adopts a one-to-many recognition method, assuming that the company employees are A, B, and C, the punch card machine needs to pre-store sample A, sample B, and sample C, and the samples here are all faces. The original features of sample A, sample B, and sample C are all 6496-dimensional, so the punching machine performs dimensionality reduction on sample A, sample B, and sample C according to the equal error rate of 0.03 to obtain feature sets (20A, 20B, 20C); Then, according to the false recognition rate of 0.000001, the sample A, sample B, and sample C are dimensionally reduced to obtain feature sets (400A, 400B, 400C). The punch card machine then stores the feature set (20A, 20B, 20C) and the feature set (400A, 400B, 400C), that is, the feature set (20A, 20B, 20C) occupies 3*20 bytes, and the feature set (400A, 400B, 400C) occupy 3*400 bytes. If employee A needs to check-in, the punch-card machine obtains the face image of employee A through the camera as a recognition sample and extracts the original features, and then uses the same method to reduce the dimensionality of the original features of employee A to obtain feature 20a and feature 400a. Before the comparison and recognition, the punching machine loads the stored feature set (20A, 20B, 20C) into the memory, and then compares 20a with 20A, 20B, and 20C respectively to obtain the following results: the similarity between 20a and 20A is 0.85, The similarity between 20a and 20B is 0.8, and the similarity between 20a and 20C is 0.75. Here, the check-in machine judges that they are the same person if the similarity is not less than 0.8. At this time, the first quantitative feature with a similarity not lower than 0.8 is (20A , 20B); the punching machine then loads the stored feature set (400A, 400B) into the memory, and then compares 400a with 400A and 400B respectively to obtain the following results: the similarity between 400a and 400A is 0.95, and the similarity between 400a and 400B It is 0.5, and the condition for the punch card machine to judge that they are the same person is that the similarity is not less than 0.9. Since only the features of sample A meet the requirement of similarity with the face image sample of employee A, the punch card machine thinks that sample A is the same person as employee A and displays it. In practical application, if the punch card machine compares 400a with 400A and 400B respectively, the following results are obtained: the similarity between 400a and 400A is 0.95, and the similarity between 400a and 400B is 0.9. The condition for judging that the punch card machine is the same person is The similarity is not less than 0.9. Then the punch card machine can judge that sample A and sample B are the same person as employee A and display it; or it can only consider that sample A and employee A with a similarity of 0.95 are the same person and display it, that is, only select the one with the highest similarity The sample is the final result; or the two similarities can be compared with the third threshold again, where the third threshold is 0.95, then the clock-in machine finally considers that sample A and employee A are the same person and displays it.

At the same time, the punch card machine can also adopt a one-to-one verification method. When the punch card machine stores the characteristics of sample A (20A, 400A), it will determine a PIN corresponding to the feature of sample A and save it to the memory of the punch card machine. middle. When employee A needs to check in, he inputs the PIN, and the punch card machine reads the features corresponding to the input PIN from the memory, that is, (20A, 400A) and loads it into the memory, and at the same time collects the face image of employee A, extracts the features and reduces the dimension to obtain (20a, 400a); the punching machine compares (20A, 20a), if its similarity is less than the first threshold 0.8, the punching fails; if the similarity is not less than 0.8, then the punching machine further compares (400A, 400a), If the similarity is not less than the second threshold of 0.9, the check-in is successful, and if the similarity is less than 0.9, the check-in fails.

In this embodiment, the terminal reduces the dimensionality of the first original feature of the first sample according to the target equal error rate and the target false recognition rate respectively to obtain the first feature and the second feature and store them; when the terminal obtains the biological image of the object to be identified After obtaining the second sample and extracting the second original feature of the second sample, the second original feature is also dimensionally reduced according to the target equal error rate and target false recognition rate respectively to obtain the third feature and the fourth feature; when the second sample In the process of comparing and identifying with the first sample, the third feature is compared with the first feature, and the fourth feature is compared with the second feature. Since the biological features of the first sample and the second sample are reduced, thus The complexity of the operation process in the comparison and identification process of the terminal is reduced, thereby accelerating the speed of terminal identification of biometric features. At the same time, the first feature and the second feature can be quantified and stored, which can effectively save storage space.

The biometric identification method in the embodiment of the present invention is described above, and the terminal in the embodiment of the present invention is introduced below. Please refer to FIG. 2. An embodiment of the terminal in the embodiment of the present invention includes:

Determining module 201, is used for determining the target equal error rate and the target false recognition rate;

The first dimension reduction module 202 is used to reduce the dimension of the first original feature of the first sample to obtain the first feature according to the target equal error rate determined by the determination module, and to obtain the first feature according to the target false recognition rate determined by the determination module. The first original feature is dimensionally reduced to obtain a second feature, and the first sample is a sample pre-stored or collected by the terminal;

A storage module 203, configured to store the first feature and the second feature obtained by reducing the dimensionality of the dimensionality reduction module;

An acquisition module 204, configured to acquire a biological image of an object to be identified to generate a second sample and extract a second original feature of the second sample;

The second dimension reduction module 205 is used to reduce the dimensionality of the second original feature acquired by the acquisition module to obtain a third feature according to the target equal error rate determined by the determination module, and to obtain a third feature according to the target false recognition rate determined by the determination module. reducing the dimensionality of the second original feature acquired by the acquisition module to obtain a fourth feature;

The identification module 206 is configured to compare and identify the third feature obtained by dimension reduction of the second dimension reduction module with the first feature obtained by dimension reduction of the first dimension reduction module and reduce the dimension of the second dimension reduction module The obtained fourth feature is compared with the second feature obtained by dimensionality reduction of the first dimensionality reduction module for recognition to obtain a recognition result.

Optionally, in this embodiment, the target equal error rate is within the first value range so that the first feature does not exceed the minimum equal error rate of the second value range, and the first value range is the target equal error rate The value range of , the second value range is the value range of the first feature;

The target false recognition rate is within the third value range so that the second feature does not exceed the minimum false recognition rate of the fourth value range, the third value range is the value range of the target false recognition rate, and the fourth The value range is the value range of the second feature.

Optionally, the identification module 206 includes:

a first judging unit, configured to judge whether the similarity between the third feature and the first feature is smaller than a first threshold;

A second judging unit, configured to judge whether the fourth feature and the second feature are smaller than a second threshold if the first judging unit judges that the similarity between the third feature and the first feature is not less than the first threshold;

The first identification unit is configured to determine that the second sample is the same as the first sample if the second judgment judges that the similarity between the fourth feature and the second feature is not less than a second threshold.

Optionally, the identification module 206 also includes:

The second identifying unit is configured to determine that the second sample is different from the first sample if the first judging unit judges that the similarity between the third feature and the first feature is smaller than the first threshold.

Optionally, the identification module 206 also includes:

The third identifying unit is configured to determine that the second sample is different from the first sample if the second judging unit judges that the similarity between the fourth feature and the second feature is less than a second threshold.

Optionally, the storage module 204 includes:

The first quantization unit is configured to quantize the first feature according to the relational expression to obtain a first quantization feature, and quantize the second feature according to the relational expression to obtain a second quantization feature;

a storage unit, configured to store the first quantization feature and the second quantization feature quantized by the quantization unit;

The recognition module 206 includes:

The second quantization unit is configured to quantize the third feature according to the relational expression to obtain a third quantization feature, and quantize the fourth feature according to the relational expression to obtain a fourth quantization feature;

The fourth identification unit is configured to compare and identify the third quantitative feature with the first quantitative feature and compare and identify the fourth quantitative feature with the second quantitative feature to obtain a recognition result.

Optionally, the relationship is:

$\mathrm{<span\; class="notranslate">\; f</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; (</span>\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; V</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; N</span>}}{{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$

Wherein, V is the feature value after dimensionality reduction of the sample, V _min is the minimum feature value after dimensionality reduction of the sample, V _max is the maximum feature value after dimensionality reduction of the sample, and the value of N is 255 or 65535.

In this embodiment, the first dimensionality reduction module 202 respectively performs dimensionality reduction on the first original feature of the first sample according to the target equal error rate and the target false recognition rate determined by the determination module 201 to obtain the first feature and the second feature, and store the The storage module 203 stores; when the acquisition module 204 acquires the biological image to obtain the second sample and extracts the second original feature of the second sample, the second dimensionality reduction module 205 also calculates the first error rate according to the target equal error rate and the target false recognition rate respectively. Dimensionality reduction is performed on the two original features to obtain the third feature and the fourth feature; when the second sample is compared and identified with the first sample, the recognition module 206 compares the third feature with the first feature, and the fourth feature is compared with the first sample. Compared with the second feature, since the biometric features of the first sample and the second sample are reduced, the complexity of the calculation process of the terminal in the process of comparing and identifying is reduced, and the speed of identifying the biometric feature of the terminal is accelerated.

Please refer to FIG. 3 for details. Another embodiment of the terminal in the embodiment of the present invention includes:

Transceiver 301, processor 302, bus 303, memory 304;

The transceiver 301, the processor 302 and the memory 304 are connected through the bus 303;

The bus 303 may be a peripheral component interconnect (PCI for short) bus or an extended industry standard architecture (EISA for short) bus or the like. The bus can be divided into address bus, data bus, control bus and so on. For ease of representation, only one thick line is used in FIG. 3 , but it does not mean that there is only one bus or one type of bus.

The processor 302 may be a central processing unit (central processing unit, CPU for short), a network processor (network processor, NP for short), or a combination of CPU and NP.

The processor 302 may further include a hardware chip. The aforementioned hardware chip may be an application-specific integrated circuit (application-specific integrated circuit, ASIC for short), a programmable logic device (programmable logic device, PLD for short), or a combination thereof. The above-mentioned PLD may be a complex programmable logic device (complex programmable logic device, CPLD for short), a field-programmable gate array (field-programmable gate array, FPGA for short), a generic array logic (generic array logic, GAL for short) or any combination.

The memory 304 may include a volatile memory (volatile memory), such as a random-access memory (random-access memory, RAM for short); the memory may also include a non-volatile memory (non-volatile memory), such as a flash memory (flash memory). memory), hard disk (hard disk drive, referred to as HDD) or solid-state drive (solid-state drive, referred to as SSD); the storage 304 may also include a combination of the above types of storage.

Optionally, the memory 304 can also be used to store program instructions, and the processor 302 calls the program instructions stored in the memory 304 to perform one or more steps in the embodiment shown in FIG. 1 , or alternative implementations thereof. , to realize the function of the terminal behavior in the above method.

The processor 302 has the following functions: determine the target equal error rate and the target false recognition rate; reduce the dimensionality of the first original feature of the first sample to obtain the first feature according to the target equal error rate determined by the determination module, and obtain the first feature according to the target equal error rate determined by the determination module The target false recognition rate determined by the determination module reduces the dimensionality of the first original feature to obtain a second feature, and the first sample is a sample stored or collected in advance by the terminal;

The memory 304 has the following function: storing the first feature and the second feature obtained by reducing the dimensionality of the dimensionality reduction module;

The transceiver 301 has the following functions: acquiring a biological image of an object to be identified to generate a second sample and extracting a second original feature of the second sample;

The processor 302 has the following functions: according to the target equal error rate determined by the determination module, the second original feature obtained by the acquisition module is dimensionally reduced to obtain a third feature, and according to the target false recognition rate determined by the determination module Dimensionality reduction of the second original feature acquired by the acquisition module to obtain a fourth feature; comparing the third feature obtained by dimensionality reduction of the second dimension reduction module with the first feature obtained by dimensionality reduction of the first dimension reduction module Identifying and comparing the fourth feature obtained by reducing the dimensionality of the second dimensionality reduction module with the second feature obtained by reducing the dimensionality of the first dimensionality reduction module to obtain a recognition result.

Optionally, the target equal error rate is within the first value range so that the first feature does not exceed the minimum equal error rate of the second value range, and the first value range is the value range of the target equal error rate , the second value range is the value range of the first feature;

The target false recognition rate is within the third value range so that the second feature does not exceed the minimum false recognition rate of the fourth value range, the third value range is the value range of the target false recognition rate, and the fourth The value range is the value range of the second characteristic.

Optionally, the processor 302 specifically has the following function: judge whether the similarity between the third feature and the first feature is less than a first threshold; if the first judging unit judges that the third feature is similar to the first feature If the similarity between the fourth feature and the second feature is not less than the second threshold; if the second judgment judges that the similarity between the fourth feature and the second feature is not less than the second threshold , then it is judged that the second sample is the same as the first sample.

Optionally, the processor 302 specifically has the following function: if the first judging unit judges that the similarity between the third feature and the first feature is less than the first threshold, then judge that the second sample is similar to the first feature. The samples are not the same.

Optionally, the processor 302 specifically has the following function: if the second judging unit judges that the similarity between the fourth feature and the second feature is less than a second threshold, then the terminal judges that the second sample is similar to the first A sample is not the same.

Optionally, the memory 304 specifically has the following functions: quantify the first feature according to the relational expression to obtain the first quantization feature, and quantize the second feature according to the relational expression to obtain the second quantization feature; The first quantization feature and the second quantization feature obtained by unit quantization are stored;

The processor 302 specifically has the following functions: quantify the third feature according to the relational expression to obtain a third quantized feature, and quantize the fourth feature according to the relational expression to obtain a fourth quantized feature; quantify the third feature The feature is compared and identified with the first quantitative feature, and the fourth quantitative feature is compared and identified with the second quantitative feature to obtain a recognition result.

Optionally, the relationship is:

$\mathrm{<span\; class="notranslate">\; f</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; (</span>\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; V</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; N</span>}}{{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$

Wherein, V is the feature value after dimensionality reduction of the sample, V _min is the minimum feature value after dimensionality reduction of the sample, V _max is the maximum feature value after dimensionality reduction of the sample, and the value of N is 255 or 65535.

In this embodiment, the processor 302 respectively performs dimensionality reduction on the first original feature of the first sample according to the target equal error rate and the target false recognition rate to obtain the first feature and the second feature and store them through the memory 304; when the transceiver 301 After acquiring the biological image of the object to be recognized to obtain the second sample and extracting the second original feature of the second sample, the processor 302 also performs dimensionality reduction on the second original feature according to the target equal error rate and the target false recognition rate respectively to obtain the third feature and the fourth feature; when the second sample is compared and identified with the first sample, the processor 302 compares the third feature with the first feature, and compares the fourth feature with the second feature, due to the reduction The biological characteristics of the first sample and the second sample are improved, thereby reducing the complexity of the calculation process of the terminal in the process of comparing and identifying, thereby accelerating the speed of terminal recognition of biological characteristics.

Those skilled in the art can clearly understand that for the convenience and brevity of the description, the specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiment, which will not be repeated here.

In the several embodiments provided in this application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or May be integrated into another system, or some features may be ignored, or not implemented. In another point, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

If the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the essence of the technical solution of the present invention or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , including several instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk, and other media that can store program codes.

As mentioned above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still understand the foregoing The technical solutions recorded in each embodiment are modified, or some of the technical features are replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (15)

1. A biometric identification method, characterized in that, comprising: The terminal determines the target equal error rate and the target false recognition rate; The terminal reduces the dimensionality of the first original feature of the first sample according to the target equal error rate to obtain a first feature, and reduces the dimensionality of the first original feature according to the target false recognition rate to obtain a second feature, and the first A sample is a sample stored or collected by the terminal in advance; The terminal stores the first feature and the second feature; The terminal obtains the biological image of the object to be identified to generate a second sample and extracts a second original feature of the second sample; The terminal reduces the dimension of the second original feature according to the target equal error rate to obtain a third feature, and reduces the dimension of the second original feature according to the target false recognition rate to obtain a fourth feature; The terminal compares and identifies the third feature with the first feature and compares and identifies the fourth feature with the second feature to obtain a recognition result. 2. The method according to claim 1, wherein the target equal error rate is within the first value range so that the first feature does not exceed the minimum equal error rate of the second value range, and the second A value range is the value range of the target equal error rate, and the second value range is the value range of the first feature; The target false recognition rate is within the third value range so that the second feature does not exceed the minimum false recognition rate of the fourth value range, and the third value range is the value range of the target false recognition rate , the fourth value range is the value range of the second feature. 3. The method according to claim 1, wherein the terminal compares and identifies the third feature with the first feature and compares the fourth feature with the second feature Recognition The recognition results include: The terminal judges whether the similarity between the third feature and the first feature is smaller than a first threshold; If the similarity between the third feature and the first feature is not less than the first threshold, the terminal judges whether the fourth feature and the second feature are less than a second threshold; If the similarity between the fourth feature and the second feature is not less than a second threshold, the terminal determines that the second sample is the same as the first sample. 4. The method according to claim 3, wherein after the terminal judges whether the similarity between the third feature and the first feature is smaller than a first threshold, the method further comprises: If the similarity between the third feature and the first feature is smaller than the first threshold, the terminal determines that the second sample is different from the first sample. 5. The method according to claim 3, wherein after the terminal determines whether the fourth characteristic and the second characteristic are smaller than a second threshold, the method further comprises: If the similarity between the fourth feature and the second feature is smaller than a second threshold, the terminal determines that the second sample is different from the first sample. 6. The method according to any one of claims 1 to 5, wherein storing the first feature and the second feature by the terminal comprises: The terminal quantizes the first feature according to a relational expression to obtain a first quantized feature, and quantizes the second feature according to the relational expression to obtain a second quantized feature; The terminal stores the first quantized feature and the second quantized feature; The terminal compares and identifies the third feature with the first feature and compares and identifies the fourth feature with the second feature to obtain a recognition result including: The terminal quantizes the third feature according to the relational expression to obtain a third quantization feature, and quantizes the fourth feature according to the relational expression to obtain a fourth quantization feature; The terminal compares and identifies the third quantitative feature with the first quantitative feature, and compares and identifies the fourth quantitative feature with the second quantitative feature to obtain a recognition result. 7. method according to claim 6, is characterized in that, described relational expression is: $\mathrm{<span\; class="notranslate">\; f</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; (</span>\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; V</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; N</span>}}{{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$ Wherein, the V is the feature value after dimension reduction of the sample, the V _min is the minimum feature value after dimension reduction of the sample, the V _max is the maximum feature value after dimension reduction of the sample, and the value of N is 255 or 65535. 8. A terminal, characterized in that, comprising: A determination module is used to determine the target equal error rate and the target false recognition rate; The first dimension reduction module is configured to reduce the dimensionality of the first original feature of the first sample to obtain the first feature according to the target equal error rate determined by the determination module, and to obtain the first feature according to the target identification determined by the determination module. Falsely reducing the dimension of the first original feature to obtain a second feature, the first sample is a sample stored or collected in advance by the terminal; A storage module, configured to store the first feature and the second feature obtained by reducing the dimensionality of the dimensionality reduction module; An acquisition module, configured to acquire a biological image of an object to be identified to generate a second sample and extract a second original feature of the second sample; A second dimensionality reduction module, configured to reduce the dimensionality of the second original feature acquired by the acquisition module to obtain a third feature according to the target equal error rate determined by the determination module, and obtain a third feature according to the determined equal error rate by the determination module. The target false recognition rate reduces the dimensionality of the second original feature acquired by the acquisition module to obtain a fourth feature; An identification module, configured to compare and identify the third feature obtained by reducing the dimensionality of the second dimensionality reduction module with the first feature obtained by reducing the dimensionality of the first dimensionality reduction module and identify the second dimensionality reduction The fourth feature obtained by the dimensionality reduction module and the second feature obtained by the dimensionality reduction module of the first dimensionality reduction module are compared and recognized to obtain a recognition result. 9. The terminal according to claim 8, wherein the target equal error rate is within the first value range so that the first feature does not exceed the minimum equal error rate of the second value range, the second A value range is the value range of the target equal error rate, and the second value range is the value range of the first feature; The target false recognition rate is within the third value range so that the second feature does not exceed the minimum false recognition rate of the fourth value range, and the third value range is the value range of the target false recognition rate , the fourth value range is the value range of the second feature. 10. The terminal according to claim 8, wherein the identification module comprises: A first judging unit, configured to judge whether the similarity between the third feature and the first feature is smaller than a first threshold; A second judging unit, configured to judge the fourth feature and the second feature if the first judging unit judges that the similarity between the third feature and the first feature is not less than the first threshold Whether it is less than the second threshold; The first identification unit is configured to determine that the second sample is the same as the first sample if the second judgment determines that the similarity between the fourth feature and the second feature is not less than a second threshold. 11. The terminal according to claim 10, wherein the identification module further comprises: A second identifying unit, configured to determine whether the second sample is similar to the first sample if the first judging unit judges that the similarity between the third feature and the first feature is smaller than the first threshold Not the same. 12. The terminal according to claim 10, wherein the identification module further comprises: A third identifying unit, configured to determine, by the terminal, that the second sample is the same as the first if the second judging unit judges that the similarity between the fourth feature and the second feature is less than a second threshold This is not the same. 13. The terminal according to any one of claims 8 to 12, wherein the storage module comprises: A first quantization unit, configured to quantize the first feature according to the relational expression to obtain a first quantization feature, and quantize the second feature according to the relational expression to obtain a second quantization feature; a storage unit, configured to store the first quantization feature and the second quantization feature quantized by the quantization unit; The identification module includes: A second quantization unit, configured to quantize the third feature according to the relational expression to obtain a third quantization feature, and quantize the fourth feature according to the relational expression to obtain a fourth quantization feature; The fourth identification unit is configured to compare and identify the third quantitative feature with the first quantitative feature and compare and identify the fourth quantitative feature with the second quantitative feature to obtain a recognition result. 14. The terminal according to claim 13, wherein the relational expression is: $\mathrm{<span\; class="notranslate">\; f</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; (</span>\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; V</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; N</span>}}{{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span>$ Wherein, the V is the feature value after dimension reduction of the sample, the V _min is the minimum feature value after dimension reduction of the sample, the V _max is the maximum feature value after dimension reduction of the sample, and the value of N is 255 or 65535. 15. A terminal, characterized in that, comprising: transceivers, processors, memory and buses; The transceiver, the processor and the memory are connected through the bus; The processor has the following functions: determining a target equal error rate and a target false recognition rate; reducing the dimensionality of the first original feature of the first sample to obtain a first feature according to the target equal error rate determined by the determination module, and reducing the dimensionality of the first original feature to obtain a second feature according to the target false recognition rate determined by the determining module, the first sample is a sample stored or collected in advance by the terminal; The memory has the following function: storing the first feature and the second feature obtained by dimensionality reduction of the dimensionality reduction module; The transceiver has the following functions: acquiring a biological image of an object to be identified to generate a second sample and extracting a second original feature of the second sample; The processor has the following functions: according to the target equal error rate determined by the determination module, the dimensionality reduction of the second original feature acquired by the acquisition module is obtained to obtain a third feature, and according to the determined The target false recognition rate is obtained by reducing the dimensionality of the second original feature obtained by the acquisition module to obtain a fourth feature; combining the third feature obtained by the second dimensionality reduction module with the first dimensionality reduction module The first feature obtained by dimension reduction is compared and identified, and the fourth feature obtained by dimension reduction of the second dimension reduction module is compared with the second feature obtained by dimension reduction of the first dimension reduction module Get the recognition result for the recognition.