JPH0312351B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to an input data recognition method for recognizing handwritten input data. In recent years, for example, translators and electronic memos,
Electronic devices that can input alphabetic characters, such as small electronic calculators that can be programmed with BASIC programs, have been commercialized. However, since this type of electronic device requires keys for each character, the number of keys increases, taking up a large amount of keyboard space, and it takes time to find the desired character key, resulting in poor operability. It was hot. On the other hand, handwriting input devices also generally exist. However, these input coordinates using electromagnetic coupling tablets, pressure-sensitive rubber, etc., which complicates the structure of the input device, increases cost, and requires a large amount of information (for example, both the X and Y coordinates are 0). ~500, etc.), it required a large amount of memory and the processing was complicated. The present invention has been made in view of the above points, and an object of the present invention is to provide an input data recognition method that enables handwritten input with a small number of input elements. DESCRIPTION OF THE PREFERRED EMBODIMENTS A case in which the present invention is implemented in a small electronic calculator will be described in detail below with reference to the drawings. FIG. 1 shows the external configuration, in which a display section 2 and a keyboard 3 are provided on the upper surface of a case body 1. The display section 2 has two upper and lower display sections 2a and 2b using, for example, a liquid crystal display element, and has a size of 5×7.
The dots allow a predetermined number of digits to be displayed such as numbers 0 to 9, alphanumeric characters A to Z, +, -, x, ÷, and arithmetic symbols. Further, the keyboard 3 has, for example, a numeric keypad, function keys, etc. arranged in a 5×6 matrix, and constitutes a handwriting input section 4. Furthermore, the keyboard 3 includes "NO" and "OFF" keys for turning the power on and off, "AC" (all clear) key, and "C" (clear) key.
key, a ``MEMO'' key for electronic memos, ``↑'' and ``↓'' keys for reading the stored contents of electronic memos, and a ``mode'' key 5 for switching between mode 1 and mode 2. Mode 1 above is a calculation mode, each key has its original function, and mode 2
In the matrix input mode, handwritten characters can be input from the handwriting input section 4. In this matrix input mode, by tracing the 5 x 6 keys of the handwriting input unit 4 with your finger, you can input numbers 0 to 9, alphanumeric characters A to Z, operation symbols +, -, ×, ÷, etc. can be entered and displayed. Then, by pressing the "MEMO" key, the character data being displayed is stored in the electronic memo. Also, characters entered by hand are recognized by the internal circuit, so for example "5+2="
If you input something like this, you can recognize these numbers and operation symbols and perform calculations. Next, the configuration of the electronic circuit provided in the case body 1 will be explained with reference to FIG. 2. Note that FIG. 2 is a functional block diagram for explaining the recognition function in the matrix input mode, and parts related to the calculation function, electronic memo function, etc. are omitted. In FIG. 2, a key input section 11 generates key input data in response to key operations on the keyboard 3 in FIG. That is, the key input section 11 generates codes for the functions of each key body in the calculation mode, and generates coordinate data of the operation keys in the handwriting input section 4 in the matrix input mode. Of course, a normal key code is generated from the pressed key, but the key code is converted and coordinate data is created and output. The coordinate data from the key input unit 11 is then sent to the input pattern memory 12 and stored as pattern data. The pattern data stored in the input pattern memory 12 is sent to a stroke number extraction block 13, where the number of strokes is extracted, and stored in the stroke number memory 14. Furthermore, the stroke feature extraction block 15 extracts stroke features, and the stroke number is extracted. It is stored in the feature memory 16. The number of strokes and stroke characteristics of the input character stored in the stroke number memory 14 and stroke feature memory 16 are sent to a matching block 17 and compared with a standard pattern sequentially read out from a standard pattern memory 18. Standard pattern memory 18 above
stores character characters in order of the number of strokes, stroke characteristics for their standard patterns, and direction characteristics for characters with the same standard pattern. The matching block 17 is connected to the stroke number memory 14.
The stroke characteristics of the standard pattern having the same number of strokes as the number of strokes written in the standard pattern memory 18
The characters are sequentially read out from the input pattern and compared with the stroke characteristics of the input pattern to perform character recognition. If there are multiple characters having the same stroke feature, the stroke feature extraction block 15 extracts the direction feature of the input pattern and stores it in the stroke feature memory 16, and the direction feature is extracted from the standard pattern memory 18. The matching block 17 again compares the directional characteristics of the read standard pattern to perform character recognition. The matching block 17 then accesses the character generator 19 in accordance with the recognized characters to generate a dot pattern and display it on the display section 1. The above-mentioned stroke characteristics indicate the direction and time order of strokes in character input.
In other words, as shown in Figure 3, the stroke direction of the X-Y coordinates is "1" for the +X direction, "3" for the -X direction, "2" for the +Y direction, and "4" for the -Y direction. stipulate. Then, the above-mentioned stroke characteristics are set for each character according to the number of strokes, and are stored in advance in the standard pattern memory 18 as a standard pattern.
For example, if it is the letter A of Alphabet, the fourth
Since it is possible to write with two strokes as shown in Figure a, and with three strokes as shown in Figure 4B, standard patterns are set for each case.
When writing A with two strokes, in the standard pattern memory 18, as shown in FIG. and its direction change are written in the stroke feature storage section 18b. This stroke characteristic storage section 18b is capable of storing two-digit data for each stroke.
Data indicating the X and Y directions are stored in the upper digits, and data indicating changes in the directions are stored in the lower digits. In the two-stroke A, in the first half of the first stroke, that is, from the bottom to the top,
Since the X coordinate is in the "1" direction and the Y coordinate is in the "2" direction, write the values in the upper digits. Also, this first
In the first half of the stroke, there is no change in direction, so
As a numerical value, write "1" in the lower digit. and,
In the second half of the first stroke, that is, from the top end to the bottom end, the X coordinate direction is the same as the first half of the stroke and there is no change, so there is no additional data. However, in the Y coordinate, the direction changes to the "4" direction, so following the "2" data, "4" is written in the first digit, and the "2" data indicating the change in direction is written in the second digit. Write. Since the first stroke ends here, a stroke end mark is written next. In the second stroke, since there is no change in the Y coordinate, data for direction "1" and direction change "1" is written only for the X coordinate. That's two strokes of A
Since the stroke characteristic for has ended, a character end mark is written following the second stroke data. In the case of A with three strokes, the number of strokes "3" is written in the stroke number storage section 18a as shown in FIG. 4b, and the stroke characteristics are set in the stroke characteristic storage section 18b in the same manner as in the above case. do. That is, in the first stroke, the X coordinate is in the "3" direction and the Y coordinate is in the "4" direction, and since there is no change in direction, their values are both "1". In the second stroke, the X coordinate is in the "1" direction, the Y coordinate is in the "4" direction, and the direction change value is "1". The third stroke is X
The coordinate is in the "1" direction, and the change value in the direction is "1". Note that there is no deviation of the Y coordinate in the third stroke. Similarly, the number of strokes and stroke characteristics of other alphanumeric characters, numbers, and symbols such as +, -, x, ÷ are stored in the standard pattern memory 18 in advance. Note that some of the above standard patterns have exactly the same pattern depending on the characters. For example, as shown in Figure 4c, the alpha alphabet “D”
In the case of the characters "P" and "P", the number of strokes and stroke characteristics are exactly the same. Therefore, for such characters, the distances between the starting points and ending points of each stroke and the next stroke are compared, and if the difference is "2" or more, a direction feature is added, and the distance is added to the standard pattern memory 18. It is stored in advance in the direction feature storage section 18c. For example, in the case of "D", as shown in Figure 4d, the starting point and ending point of the first stroke and the second stroke are at the same position, so the direction feature is
The coordinates and Y coordinates are all "0". Further, in the case of "P", as shown in FIG. 4e, the first stroke and the second stroke have starting points at the same position, but end points at different positions. That is, since the end point of the second stroke is away from the end point of the first stroke in the Y coordinate only in the "2" direction, "2" is set as the direction feature only in the Y coordinate of the second stroke. The other X, Y coordinates of the first stroke and the X coordinate of the second stroke are "0". Furthermore, if the character is one stroke, the direction feature is set from the distance between the starting point and the ending point of the stroke. Furthermore, in the standard pattern memory 18,
A flag memory is provided for each character character, and "1" is stored in the flag memory of a standard pattern whose number of strokes matches the input data. In addition, since you are tracing the top of the key with your finger,
You end up pressing multiple keys at once. Therefore, in this embodiment, three
The direction feature is detected only when more than 1 keys are pressed. Next, the operation of the above embodiment will be explained. When performing calculations, mode 1 is specified using the "mode" key 5, but the key input process in this case is the same as the conventional one, so a description of its operation will be omitted.
When performing handwritten input, mode 2 is designated using the "mode" key 5 to set the matrix input mode. Then, the user inputs characters by tracing the keys provided in the handwriting input unit 4, that is, the keys arranged in a 5×6 matrix, with a finger. For example, suppose that the character A is input using three strokes like 〓〓. Figure 5a shows the order of key operations in the first stroke when the letter A is traced with a finger, b shows the order of key operations in the second stroke, and c shows the order of key operations in the third stroke. This is an example. Coordinate data is outputted from the key input section 11 in FIG. 2 by the above key operations and written into the input pattern memory 12. For example, in the first stroke shown in FIG. has coordinate data [3,
2] is output. The above coordinate data is 1 with the last stroke end mark of each stroke and the last character end mark of the last stroke.
It is written into the input pattern memory 12 in units of strokes. The pattern data written in the input pattern memory 12 in this case is shown in Table 1 below.

【table】

[Table] The above stroke end mark is written when continuous key input is interrupted, and the character end mark is written when there is no next key input for a predetermined period of time, indicating that the input of that character has ended. Thereafter, input character recognition processing is performed according to the flow shown in FIG. First, as shown in step _S1 in FIG. 6, the input pattern held in the input pattern memory 12 is sent to the stroke number extraction block 13, and the number of strokes is extracted by counting the stroke end mark and character end mark. and written into the stroke number memory 14. In addition, the number of strokes extraction block 13
The input pattern given to is further sent to the stroke feature extraction block 15, and the step shown in FIG.
As shown in step _S2 , X-direction stroke feature extraction is performed, and Y-direction stroke feature extraction is performed in step _S3 . X in step _S2 above
Details of directional stroke feature extraction will be described later. Stroke feature extraction in the X direction and Y direction shown in steps S ₂ and S ₃ is repeated for each stroke feature of the input character, and the extraction results are written into the stroke feature memory 16. after that,
Proceed to step _S4 and perform stroke number matching. That is, the matching block 17 compares the number of strokes of the input character held in the number of strokes memory 14 with the number of strokes of each character stored in the standard pattern memory 18, and the matched characters are stored in the standard pattern memory. Flag “1” in flag memory in 18
stand up. Next, the process proceeds to step _S5 , where the stroke features of the characters for which the flag "1" is set in the flag memory are sequentially read out from the standard pattern memory 18, and are compared with the contents held in the stroke feature memory 16 to find characters with matching stroke features. . The details of the stroke feature matching in step _S5 will be described later. After the above matching process is completed, the process proceeds to step _S6 .
In _S5 , it is determined whether the number of matched characters is one or two or more. If there is only one character, that character is determined as the input character, and the character recognition process ends; however, if there are two or more characters, the process proceeds to step _S7 to extract the directional features of the starting point and end point, and then to step _S8 . The input character is finally determined by performing feature matching on the starting point and ending point.
The details of step _S7 will be described later. After recognizing the input characters as described above, the character generator 19
is accessed to generate a dot pattern and display it on the display section 1. Next, details of the X-direction stroke feature extraction in step _S2 will be explained with reference to the flowchart of FIG. When performing stroke feature extraction in the X direction, the matching block 17 first specifies the data of the first stroke of the input pattern by setting the stroke number S to "1", and then sets the initial value shown in step _X1 in FIG. Configure settings. That is,
The value of a key operation order counter N that is incremented for each stroke including the stroke end mark and character end mark, and the value of a counter L corresponding to this N are each set to "3",
A counter M that counts stroke characteristics for each stroke is set to "1". _{Next ,} proceed to step The coordinate K _N-1 is subtracted and the difference value D _N is stored in the register D. Then, as shown in step _X3 , the value of the difference value D _N is stored in the corresponding part of the register D' as the difference value D' _L. These steps X ₂ and X ₃
The process in step X2 and _X3 is repeated in each step, with the value of N ranging from the initial value " ₃ " to the value of N corresponding to the final coordinate data of the currently specified stroke. Sometimes registers D and D' store all difference values D _N and D' _L for that stroke. after that,
Proceed _{to step} Decide whether or not. As a result of this judgment, D′ _L-1 is “0”
If so, in _step
The value of L is set to "-1", and the value of L becomes "2" at step X ₆ .
Determine whether or not. As a result of this judgment, if the value of L is not "2", the process returns to step _X4 and the same process is repeated. That is, by processing steps X ₄ to X ₆ , the difference value for the current Nth X coordinate is calculated.
The process goes back until the difference value D′ _L-1, which is different from D _N , is obtained. _{Then ,} if _{it is} determined _{in step} It is determined whether the difference values D′ _L-1 are the same. If D _N and D′ _L-1 are the same at step X ₇ , it means that the key has been pressed three times in _the It is determined whether the difference value D _N for the Nth X coordinate is smaller than "0". Here, if D _N <0 _, it means that the X coordinate has changed in the direction of "3", so in _step If so, it means that the X coordinate has changed in the direction of "1", and "1" is written in C' at step _X10 . Step X ₁₁ after executing the above step X ₉ or X ₁₀
Then, it is determined whether or not the stroke feature extraction in the current stroke pattern data is the first. That is, it is determined whether the value of the counter M that counts the number of features in each stroke is "1" or not,
If NO, in step _X12 it is determined whether the previous stroke feature C _SM-1 and the current stroke feature C' are the same. And stroke characteristics
If C' and C _SM-1 are not equal, or if it is determined that M=1 in step _X11 , proceed to _step It is stored in the Mth position of the register C _S that stores stroke characteristics. Then, in step _X14 , after adding ₁ to the value of M, proceed to step
Add "+1" to the value. It should be noted that the process also proceeds to step _X15 when the determination results at steps _X6 and _X12 are YES and when the determination result at step _X7 is NO. When _{the process of step X15 is completed, the value of N is stored in L in step X16 , and the} process proceeds to _step
Proceed to and determine whether K _N is a character end. If it is determined at step _X18 that the character is not at the end, the process returns to step _X1 and the same process is repeated. Also, in step X ₁₇ above
When it is determined that _KN is the stroke end, proceed to step _X19 , specify the pattern data for the next S+1st stroke, and then
Return to X ₁ and repeat the same process. By repeating these processes and reading out the last coordinate data of the last stroke to extract the stroke features, the step
At _X18 , _KN is determined to be the character end, the operation for extracting the stroke feature in the X direction is completed, and the process proceeds to step _S3 in FIG. Although the details of the Y-direction stroke feature extraction in step _S3 of FIG. 6 are not shown, it is almost the same as the case of the X-direction shown in _FIG . In addition, "2" and "4" are written as stroke characteristics in the Y direction at steps corresponding to steps _X9 and _X10 in FIG. 7, respectively. Next, the stroke feature matching process shown in step _S5 of FIG. 6 will be explained with reference to the flowchart of FIG. First, as shown in step _Y1 , the standard pattern memory 18 is accessed,
Check whether the storage flag F _N of the flag memory is "1". The above F _N is a stroke number determination result flag of the Nth character of the standard pattern. The value of N is initially set to "1" and is checked from the standard pattern stored at address 1 in the memory 18. And "1" in flag F _N
If it is, the process proceeds to step Y ₂ , where the stroke features of the input character held in the stroke feature memory 16 are input to the matching block 17, and compared with the stroke features read from the standard pattern memory 18 in step Y ₃ . do. Then, it is determined in step _Y4 whether or not the above two stroke characteristics match, and if they do not match, "0" is written in the flag _FN as shown in step ^Y5 .
Then proceed to step _Y6 . Further, if it is determined in step ^Y4 that the stroke characteristics match, or if it is determined in step _Y1 that the flag _FN is not "1", the process proceeds to step _Y6 . At step _Y6 , the value of N is incremented by "+1", and the process then proceeds to step _Y7 , where it is determined whether the value of N has become larger than the number of characters M of the standard pattern. If the value of N is not greater than M, the process returns to step _Y1 and the same process is repeated. That is,
The number of strokes flag F _N in the standard pattern memory 18 is “1”
Only the stroke features of the characters that are set are compared with the stroke features of the input character, and if they do not match, the flag F _N is set to "0". When access to all characters written in the standard pattern memory 18 is completed, the determination result in step _X7 becomes YES, and the stroke feature matching process is completed. Next, details of extraction of the direction feature of the start point and end point in step _S7 of FIG. 6 will be explained with reference to the flowchart of FIG. In step _S7 above, character recognition is achieved by matching stroke features.
This is executed when the character is not determined, and first, as shown in step _Z1 of FIG. 9a, it is determined whether the number of strokes of the recognized character is one. If the number of strokes is two or more, the process proceeds to the process flow shown in FIG. 9b, but if it is one stroke, the process proceeds to the next step _Z2 in the flow shown in FIG. 9a. This step
For Z ₂ , _{the X} coordinate of the stroke end point of the input character K ^E
Subtract _{the X} coordinate of the stroke start point K ^S
1, 0, 1”. If the subtraction result is "-2" or less, this indicates that the X coordinate on the coordinates shown in Figure 3 has deviated by "3" or more between the start and end of the stroke, so the subtraction result is shown in step _Z3 . The direction feature "3" is determined as follows. Further, if the above subtraction result is "2" or more, it indicates that the direction of the stroke is in the direction of the X coordinate "1", so the direction feature "1" is determined as shown in step _Z4 . Also, the above subtraction result is "-1, 0, 1"
If so, the moving distance on the X coordinate of the stroke is "2" or less, so in this case step Z ₅
The direction feature “0” is determined as shown in . After determining the directional characteristics of the X coordinate, proceed to step _Z6 , and subtract the Y coordinate ^KSY of the _stroke start point from the _Y coordinate ^KEY of the stroke end point, and whether the subtraction result is less than or equal to "-2". , "2" or more, or "-1,0,
1", and determine the direction feature of the Y coordinate in the same way as for the X coordinate. That is, if the above subtraction result is less than or equal to "-2", the direction feature is set to "4" at step _Z7 ; if the subtraction result is greater than or equal to "2", the direction feature is set to "2" at step _Z8 , and the subtraction result is set to "-1". , 0, 1'', the direction feature ``0'' is determined in step <sub>Z9</sub> . This completes the extraction of the direction feature of the start point and end point when the input character is one stroke, but if the input character is two or more strokes, proceed from step Z <sub>1</sub> to Figure 9b.
Processing is performed according to the flow shown in . First, in step <sub>Z11</sub> , the Nth character of the input character is
Subtract the X coordinate K <sup>SN -1</sup> <sub>X of</sub> the previous stroke start point from the X coordinate K SN
1”. That is, step Z <sub>11</sub> above
Find the difference between the start positions of the first and second strokes. If the above subtraction result is less than or equal to "-2", the direction feature is set to "3" in step <sub>Z12</sub> , and if the subtraction result is greater than or equal to " <sub>2</sub> ", the direction feature is set to "1", "-1, 0," 1”, step
Determine the direction feature "0" at <sub>Z14</sub> . Next, proceed to step Z <sub>15</sub> and move to the end point of the Nth stroke, X.
From the coordinate K <sup>EN</sup> <sub>X</sub> to the X coordinate of the previous stroke end point
E <sup>EN-1</sup> Subtract <sub>X</sub> and check if the subtraction result is “-2” or less,
Determine whether it is "2" or more or "-1, 0, 1". That is, the difference between the positions of the end points of the first stroke and the second stroke is determined. If the above subtraction result is less than or equal to "-2", the direction feature is set to "3" in step <sub>Z16</sub> , and if the subtraction result is greater than or equal to "2", the direction feature is set to "1" in step <sub>Z17</sub> , and the subtraction result is set to "-1". , 0, 1'', the direction feature ``0'' is determined. After that, proceed to step Z ₁₉ and set the Y coordinates (K ^SN _Y , K ^SN-1 _Y ), (K ^EN _Y ,
For K ^EN-1 _Y ), the direction feature is determined in the same manner as for the X coordinate above. After determining the direction characteristic of the Y coordinate, the process proceeds to step _Z20 , where it is determined whether or not the process up to the final stroke has been completed.If the process has not been completed, the process returns to step _Z11 ; The direction feature extraction of the starting point and ending point in the above case is completed. Thereafter, the process proceeds to step _S8 in FIG. 6, where directional feature matching of the starting point and ending point is performed. Furthermore, in the above embodiment, characters are entered by directly tracing the keys provided on the handwriting input section 4 with a finger, but a sheet is placed over the handwriting input section 4, and characters are input using a finger or a pen on top of the sheet. You can also write Further, in the above embodiment, keys used for ordinary calculations and keys for handwriting input are used, but it is of course possible to use keys exclusively for handwriting input. As described above, according to the present invention, it is possible to provide an input data recognition method that can reliably recognize handwritten input even if the data input section is configured with a small number of input elements.

[Brief explanation of the drawing]

The drawings show one embodiment of the present invention; FIG. 1 is an external front view, FIG. 2 is a functional block diagram, FIG. 3 is a diagram showing stroke directions on coordinates, and FIGS.
5e is a diagram showing an example of a standard pattern stored in the standard pattern memory; FIGS.
FIG. 6 is a flowchart showing the input character recognition operation, FIG. 7 is a flowchart showing details of the X-direction feature extraction step in FIG. 6, and FIG.
A flow chart showing the details of the stroke feature matching step in Figure 9, Figures 9a and 9b are
2 is a flowchart showing details of the step of extracting directional features of the starting point and ending point in the figure. 1...Case body, 2...Display section, 3...Keyboard, 4...Handwriting input section, 11...Key input section, 12...Input pattern memory, 13...Number of strokes extraction block, 15...Stroke characteristics Extraction block, 17... Matching block, 18... Standard pattern memory.

Claims (1)

[Claims] 1. A data input section formed by arranging input elements in a matrix, means for extracting the number of strokes of data input from this data input section, and dividing the data input section into four directions of X-Y coordinates, at least means for extracting stroke characteristics in a certain direction when three input elements are inputted in that direction; a standard pattern memory that stores the number of strokes, stroke characteristics, and direction characteristics of each standard pattern; and the number of strokes means. A stroke number matching means that matches the number of strokes of the extracted input data with the number of strokes of the standard pattern stored in the standard pattern memory and detects a standard pattern with a matching number of strokes; On the other hand, a stroke feature matching means matches the stroke features with the stroke features extracted by the stroke feature extraction means and detects a standard pattern with matching stroke features, and the stroke feature matching means combines the recognized characters into one. An input data recognition method characterized by comprising a means for extracting the directional features of the start point and end point of the input data when the narrowing cannot be narrowed down, and performing matching with the directional features of a standard pattern to finally determine the input character. .