CN104808701A - Cloud vision-based automatic eye make-up apparatus - Google Patents

Abstract

Automatic eye makeup machine device based on cloud vision, based on the control mechanism, camera, LED fill light, cloud control chip loaded with recognition and eye tracing algorithms, microcontroller, wireless transmission module, USB interface, eye tracing brush , eye clipper, stepping motor and other equipment, designed and implemented an automatic eye makeup machine device based on cloud vision. This invention can solve the time-consuming and unprofessional effects of human eye makeup. It can be set for multiple uses once or multiple times, and the use mode can be changed as needed. It satisfies the terminal needs of beauty lovers and can eventually be promoted all over the world, thus becoming a necessity of life and generating huge economic and social benefits.

Description

Automatic eye makeup machine device based on cloud vision

technical field

The invention belongs to the field of small household appliance manufacturing, robot and makeup technology, and relates to an automatic eye makeup machine device based on cloud vision.

Background technique

Makeup has a long history in the world, whether in the film and television industry or in daily life, it has become a common daily behavior for most people. Moreover, with the development of society, more and more attention is paid to the experience of makeup technology. The profession of professional makeup artists has also been paid more and more attention by people, but most people have neither the time nor the money to hire professionals to make up for themselves. In the whole facial makeup, the eyes are the window to the soul, and the big and spiritual makeup that presents the eyes is definitely a successful beauty makeup.

As the saying goes: what suits you is the best, and this sentence also applies to eye makeup. Everyone's face shape is unique, we can't just care about the good-looking painting, we must find the eye makeup that suits us. The general steps of eye makeup are as follows:

1. First of all, the base makeup. The base makeup is absolutely not to be ignored for a perfect makeup look. There are many ways to apply the liquid foundation. It is recommended to use a medium-sized brush to sweep it slowly to make it more delicate.

2. Immediately afterwards, shade the eyelids to make the face more three-dimensional and compact.

3. Then we start the part of the eye makeup, first of all, use a large area of the eye socket as a base, and use natural colors.

4. Then gently sweep a layer of light green eye shadow within the range marked in the picture, focusing on the area around the eyeliner.

5. Next, use a small eyeliner stick to pick up a little eye shadow and gently sweep it away on the lower eyeliner.

6. Immediately afterwards, swipe with a deeper layer of dark green eye shadow within the range marked in the picture, especially focusing on the part at the end of the eye.

7. Then there is the eyeliner. The eyeliner is in the form of a half package, the upper eyeliner is drawn in a smooth way, and the lower eyeliner extends from the eyeball part to the back, and is connected with the upper eyeliner at the triangle of the corner of the eye.

Finally, sweep a light layer of pink blush on the smiling muscles to complete the whole makeup.

In order to achieve the effect of a professional makeup artist, and to facilitate people to save time and achieve professional makeup technology. The present invention has developed an automatic eye makeup machine device based on cloud vision. First, there must be a recognition device that can analyze the user's face shape. Therefore, the facial features recognition algorithm used in face recognition is inside the chip of the device, and several eye-shaped design schemes are set in advance, which can be displayed on the screen through a USB interface or a wireless transmission module. And based on empirical calculations, what kind of eye makeup and makeup color is suitable for this face shape, etc., can also be defined and selected by yourself. The camera is used to extract the location of the user's eyes. The width of the automatic eye makeup machine can be set to different models, and the user can stick to the face. It consists of: a camera, a cloud control chip loaded with recognition and eye makeup algorithms, a wireless transmission module, a USB interface, an eyeliner brush, an eyeliner pen, a horizontal stepping motor, and a vertical stepping motor. That is to say, it is composed of microcontroller, sensor module, power module, drive module, support unit, eye makeup unit and so on.

Contents of the invention

The purpose of the present invention is to provide an automatic eye makeup machine device based on cloud vision.

Above-mentioned purpose realizes by following technical scheme:

Automatic eye makeup machine device based on cloud vision, based on the control mechanism, camera, LED fill light, cloud control chip loaded with recognition and eye tracing algorithms, microcontroller, wireless transmission module, USB interface, eye tracing brush , eye clipper, stepping motor and other equipment, designed and implemented an automatic eye makeup machine device based on cloud vision. This invention can solve the time-consuming and unprofessional effects of human eye makeup. It can be set for multiple uses once or multiple times, and the use mode can be changed as needed. It satisfies the terminal needs of beauty lovers and can eventually be promoted all over the world, thus becoming a necessity of life and generating huge economic and social benefits. Its feature is to place the device in front of the face, capture the face through the camera, analyze the collected image, analyze the eye shape of the suitable face through the built-in algorithm, and manually confirm it through the display screen or mobile terminal; Put it on the eye and the eye, the camera will locate the existing eye hair position, and then press the confirmation button to start the smart eye tracing machine to trace the eye. The specific process is: the camera collects information and transmits it to the algorithm chip. After the microcontroller receives the positioning information, it drives the motor to fine-tune the positioning of the pen. The eye pen moves repeatedly to the set eye shape according to the position information. Deeper, the effect of different shades can be formed through the frequency of movement, thereby forming a sense of hierarchy.

  The interface of the present invention may include a display and a man-machine operation interface, and the display may be a mobile phone, a tablet type touch control or non-touch control. The user can simulate and draw the eye shape through the display preset to determine whether it is what the user needs, or set the eye shape by himself.

This intelligent eye tracing machine is equipped with an abnormal interrupt button. If there is any abnormality, you can press the button to make an emergency stop to avoid abnormal situations.

This intelligent eye tracing machine is equipped with a pincer nose fixing device, which can fix the device on the upper part of the face like wearing glasses, or use the head positioning piece to assist positioning, or lightly press the fixing device with your hands.

The steps of using this smart eye tracing machine are as follows:

1. Start the device, the camera will take a global snapshot of the face picture, and then display it through the display terminal of the device, and upload the picture to the cloud through wireless, analyze the user's facial features through the cloud algorithm module, and intelligently select and The mode judged by experts partially replaces the eye shape suitable for the face shape, displays it to the user, and sends the selected eye shape to the device storage area.

2. After fixing the device on the head, use the camera to precisely locate the position of the user's eye hair. The positioning can accurately locate the eye hairs through edge detection and color difference distinction. It is only on the eye, so there is no other interference, so the visual positioning method is simple and reliable.

3. After analyzing the three points of the eye starting point, eye peak, and eye end through image processing combined with the three-point ratio and eye peak protrusion, start from the eye peak, control the motor to drive the eye tracing mechanism to push the eye pencil repeatedly according to the position information Paint, and set empirical values for the step size and frequency through the frequency converter, so that three-dimensional and layered effects can be obtained.

4. Pull out the excess eye hair by driving the eye clipper by the motor.

The eyebrow drawing algorithm in the cloud is mainly described as follows: how to obtain different types of features or attributes:

1. Geometric attributes, focusing on calculating shape features through eyebrow coordinates.

2. Holistic: Each individual image is cropped into a standard region before applying PCA (Principal Component Analysis) to a set of available samples. Mainly (i) discard as many background pixels as possible (ii) define the same part for all frontal faces and side faces, including the most prominent facial features (forehead, nose, mouth and chin, etc.) (iii) as much as possible Remove the hair from the profile image. Contrastive normalization should also be used for processing images taken in uncontrolled environments. The final cropped image is converted to a grayscale image and represented as a 1D concatenated image row vector.

3. Texture description, summarizing the texture features of the entire image or small areas around the coordinates.

All geometric attributes and most texture attributes originate from the location of keypoints in the brow image. For the image, 26 keypoints have been automatically calculated using the Active Shape Model (ASM) technique.

These descriptors summarize the texture features of the entire image or the coordinates of the surrounding area.

Rotation-Invariant Co-occurrence of Neighboring LBPs (RIC-LBP) is an extension of classical Local Binary Patterns (LBPs) aimed at improving illumination invariance and, as the name suggests, rotation invariance. In preliminary tests, RIC-LBP outperformed other extensions of LBPs. For each grayscale image, feature vectors of size 136 are extracted.

Gabor filters are widely used in edge detection and texture analysis. Each coordinate is computed through the center of a window whose normalized width is 15% of the distance between two anchor points. It is found through experiments that this width can optimize the accuracy of RIGF features.

SIFT (Scale Invariant Feature Transform) is a popular local image descriptor, in fact, it has become a standard in computer vision such as object recognition, image registration, and content-based image retrieval. Therefore, SIFT-based descriptors have been used to characterize the neighborhood of the extracted coordinates in independent images. SIFT acts on monochrome images, uniformly invariant to scaling and rotation, and partially invariant to affine distortion and illumination changes. Recently, there have been different types of color descriptor-based algorithms that apply SIFT algorithms to different color schemes, which are aimed at improving illumination invariance and discrimination ^] . The present invention uses the following color descriptors:

1) Opponent-SIFT, an opposition-based color space obtained through RGB color space transformation . , , .

2) C-SIFT, based on the color invariance model and the reflection spectrum characteristics of the color mechanism, the first two parts can basically be represented by the standard opponent-SIFT color space, . The third part is the intensity channel.

3) rg-SIFT, which consists of two SIFT descriptors r and g, the chromaticity of r and g can be obtained from the RGB color model, .

4) rgb-SIFT is a SIFT descriptor independently calculated by RGB channels.

C-SIFT and rg-SIFT are intensity invariant, while opponent-SIFT and rgb-SIFT are also invariant to the color change of the light source in different images. All color descriptors are evaluated as separate SIFT descriptors of size 128 for associating the corresponding color model for each channel into simple vectors. Through preliminary experiments on the dataset, the color descriptor outperforms the standard SIFT. Considering each color descriptor individually, C-SIFT provides the best accuracy. When a color descriptor is grouped with other attributes, it turns out that C-SIFT is again slightly better, therefore, C-SIFT is chosen as the reference descriptor for image experiments. Moreover, C-SIFT is only intensity invariant, which makes it particularly suitable for processing images acquired in controlled environments. When dealing with images that have fewer or no constraints when taken, such as images, the descriptor provides higher invariance, which leads to better results.

         

Description of drawings:

Figure 1 is an automatic eye makeup machine device based on cloud vision, 1) eye tracing mechanism 2) vertical stepping motor 3) horizontal stepping motor 4) camera 5) power module 6) nose pincer 7) algorithm chip module 8) wireless Transmission module 9) Usb interface

10) Inverter

Figure 2 is a diagram of the eye tracing mechanism, 11) eye powder tank 12) eye brush 13) eyelash curler

Beneficial effect :

1) Save costs, replace manual work, and achieve professional results;

2) Intelligently analyze facial features and choose different eye shapes;

3) The virtual effect display can also be selected once and used repeatedly;

4) High-precision positioning, small particles, strong layering;

5) The equipment is compact, easy to carry, easy to use, and available anytime and anywhere;

6) Intelligent equipment, leading the industry leader. The future society will be an era of smart cities, where many jobs can be replaced by smart machines. The research and development of this product will cause a new revolution;

Detailed ways:

Above-mentioned purpose realizes by following technical scheme:

1) An automatic eye makeup machine device based on cloud vision, which consists of: a control mechanism-based system, a camera, LED fill light, a DSP chip loaded with recognition and eye tracing algorithms, a microcontroller, a wireless transmission module, and a USB interface , an eye-tracing brush, an eye clipper, and a stepping motor device, the design and implementation of an automatic eye makeup machine device based on cloud vision.

2) According to claim 1, the automatic eye makeup machine device based on cloud vision is characterized by a board composed of a DSP chip loaded with recognition and eye tracing algorithms, a microcontroller, a wireless transmission module, a USB interface, and a power interface , the stepper motor is connected to the microcontroller, the LED fill light is connected to the camera, the camera is connected to the microcontroller, the eyebrow clamp is connected to the stepper motor, the eyebrow pencil core is connected to the stepper motor, the frequency converter is connected to the stepper motor connected.

3) According to claims 1 and 2, the automatic eye makeup machine device based on cloud vision is equipped with an abnormal interrupt button, and if there is any abnormality, it can be stopped by pressing the button to avoid abnormal situations.

4) According to claims 1, 2, and 3, the intelligent eye tracing machine is equipped with a pincer nose fixing device, which can fix the device on the upper half of the face like wearing glasses, and can also use the head positioning piece to assist in positioning .

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention should be included in the protection of the present invention. within range.

Claims (4)

1. based on the automatic eye adornment machine of cloud vision, its composition comprises: system based on control gear, camera, LED light supplement lamp, loading identification and retouch the dsp chip of an algorithm, microcontroller, wireless transport module, USB interface, eyeliner brush, folder eye device, stepper motor equipment forms, and designs the automatic eye adornment machine achieved based on cloud vision.

2. according to claim 1, based on the automatic eye adornment machine of cloud vision, it is characterized in that by loading the dsp chip identifying and retouch an algorithm, microcontroller, wireless transport module, the board of USB interface and power interface composition, stepper motor is connected with micro controller, and LED light supplement lamp is connected with video camera, video camera is connected with micro controller, folder eyebrow device is connected with stepper motor, and pen core of penciling one's eyebrows is connected with stepper motor, and frequency converter is connected with stepper motor.

3., according to claim 1 and 2, based on the automatic eye adornment machine of cloud vision, being provided with abnormal break key, if any any exception by pressing one-key emergent stop, avoiding abnormal conditions to occur.

4. according to claim 1,2,3, this intelligence is retouched a machine and is provided with pince-nez stationary installation, equipment can be fixed on the first half of face as wearing glasses, also can carry out auxiliary positioning by head restraint.