JPH0211950B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (1) Description of the field to which the invention pertains The present invention relates to an online handwritten character recognition system in which character recognition ability does not deteriorate even due to incorrect stroke order or number of incorrect strokes.

(2) Description of conventional technology Unlike OCR, in conventional online handwritten character recognition methods, the input pattern to be recognized is: The method used was to perform recognition by matching only this pattern with a standard pattern with the same number of strokes. In this case, there was a restriction that the number of strokes had to be written down correctly. However, the more complex the shape, such as a kanji with a large number of strokes, the more likely it is that the number of strokes is incorrect, and the constraint of maintaining the correct number of strokes was a heavy burden on scribes.

One way to solve this problem is to register all the various ways of writing characters with different numbers of strokes as standard patterns for characters that are likely to have incorrect stroke counts. However, it is practically impossible to list all the different ways of writing each character with different numbers of strokes, and even if only some of them were registered as standard patterns, the number of standard patterns would be enormous. However, the calculation process required for recognition increases rapidly, and if the number of incorrect strokes is not registered, it will always be recognized incorrectly.

Another solution is to register only the correct stroke order and number of strokes as the standard pattern, and even if the input pattern is written in the wrong stroke order or wrong number of strokes, the pattern with the smaller number of strokes can be registered as the standard pattern. It was thought that recognition could be performed by making all strokes correspond to the strokes of the opponent's pattern.
56-160654), but if there is an inclusive relationship between the input pattern and the standard pattern, one will be embedded in the other, causing misrecognition.If the strokes of the smaller number of strokes are superimposed, the incorrect number of strokes will be written. This method has drawbacks such as the inability to deal with the unique handwritten deformation that occurs when

(3) Purpose of the invention The present invention solves the above-mentioned conventional problems, and
By simply registering only the correct stroke order and number of strokes as standard patterns, writing in the wrong stroke order and number of strokes is allowed, and in particular, handwriting deformation caused by writing with the wrong number of strokes can be dealt with without degrading character recognition ability. The purpose is to provide an online handwritten character recognition method.

(4) Description of structure and operation of the invention The details of the present invention will be explained below with reference to Examples.

FIG. 1 is a functional block diagram of a device used in one embodiment of the present invention, in which 1 is a character information input device;
2 is a preprocessing device, 3 is a stroke feature point extraction device, 4 is a standard pattern storage device, 5 is an inter-stroke distance calculation device, 6 is a stroke correspondence device,
7 is a stroke feature point resynthesis device, 8 is an inter-pattern distance calculation device, and 9 is a minimum distance value detection device.The operation of each device will be specifically explained below.

The character information input device 1 is composed of an existing tablet, and inputs the X and Y coordinates of the handwriting of each stroke of the input pattern at regular intervals, and sends a series of the strokes to the preprocessing device 2.

The preprocessing device 2 performs noise removal, smoothing, normalization of position and size, etc. on the line of pen points sent from the character information input device 1 based on known techniques, and converts the resulting line of pen points into strokes. Feature point extraction device 3
Send to.

The stroke feature point extracting device 3 selects pen points at equal intervals with a predetermined step distance value from the pen point string of each stroke sent from the preprocessing device 2, and extracts a feature point string representing each stroke. Extract.

The input pattern is denoted as S〓, and S〓 consists of N strokes S _{1 ,} ..., Si, ... _SN . _1i ,
..., x _oi , y _oi ), and if the predetermined increment distance value is Δ and the squared distance is used, for example, (x _ki −x _k+1 , i) ² + (y _ki − y _{k+ 1} , i) ² = Δ ² (
1
≦k≦n-1). Here, n is the number of feature points included in Si, and i represents handwriting with 1≦i≦N, which may be different from the correct stroke order.

The stroke feature point sequence extracted from the input pattern S is sent to the inter-stroke distance calculation device 5 and the stroke feature point resynthesis device 7.

The standard pattern storage device 4 stores the X and Y coordinates of stroke feature points extracted at equal intervals using the above-mentioned increment distance value Δ from each average pattern written in the correct stroke order and number of strokes for all character categories to be matched. The stroke feature point sequence that stores the value series and has the correct stroke order and number of strokes is sent as a standard pattern to the interstroke distance calculation device 5 and the stroke feature point recombination device 7.

The inter-stroke distance calculation device 5 calculates all strokes between the feature point sequence of the input pattern S sent out from the stroke feature point extraction device 3 and the feature point string of the standard pattern sent out from the standard pattern storage device 4. Calculate the inter-stroke distance for the pair.

The standard pattern is denoted as R〓 _, and R〓 consists of _M strokes R ₁ , ...Rj, ...R ,
y′ _1j , ..., x′ _nj , y′ _nj ). Here, m is the number of feature points included in Rj, and j is 1≦
j≦M represents correct Bijun. i-th stroke Si of input pattern S〓 = (x′ _1i , y′ _1i , ..., x′ _oi , y′ _oi
)
(1≦i≦N) and the j-th stroke Rj = (x′ _1j , y′ _1j , ..., x′ _nj , y′ _nj ) of the standard pattern R〓, the inter-stroke distance d _ij is, for example, the squared distance When using d _ij =α _ij・_p 〓 ^k=1 {(x _ki −x′ _kj ) ² +(y _ki −y′ _kj ) ² }, it is calculated. Here p=
min (m, n), and α _ij is a factor that evaluates the difference in the number of feature points, for example, α _ij = max (m, n)/min
(m, n) is used. d _ij is calculated for all stroke pairs, and a total of M·N inter-stroke distance values are sent to the stroke matching device 6.

The stroke matching device 6 uses the M/N inter-stroke distance values sent from the inter-stroke distance calculating device 5 and the stroke order information of the input pattern and the standard pattern to calculate the omission between the input pattern and the standard pattern. This is a device that determines stroke correspondence without a stroke, and the processing is divided into two stages.

In the first stage of processing, the number of strokes M, N
When M≧N, the corresponding stroke of the i-th stroke Si of the input pattern is taken as the ji-th stroke of the standard pattern, and _N 〓 ⁱ⁼¹ dij _i is the minimum and j ₁ , j ₂ ,..., Determine N stroke pairs that satisfy the condition that there is no overlap in j _N , and when M<N, the corresponding stroke of the j-th stroke R _j of the standard pattern _is
Determine _M stroke pairs with the minimum stroke M 〓 ^j=1 dij _j and no overlap among i ₁ , i ₂ , . . . , i _M .

In the second stage of processing, the 1M-N1 strokes included in the pattern with the largest number of strokes that remained uncorresponded in the previous step are the same as the corresponding strokes of the previous pattern that are closest in stroke order. When M≧N, j ₁ , ..., j _N determined in the previous step are rearranged so that j′ ₁ <j′ ₂ <
...<j′ _N , then if j representing the stroke order of uncorresponding stroke Rj of the standard pattern is j<j′ ₁ , it is the same as Rj′ ₁ , j′ _k <j<j′ _k+1 (1 ≦k≦N-1
)
Then, it is the same as R _j ′ _k , if j′ _N < j, it is the same as R _j ′ _N , and when M < N, i ₁ determined in the previous step,
…, i _M are rearranged and set as i′ ₁ < i′ ₂ <…< i′ _M. If i representing the stroke order of uncorresponding strokes Si of the input pattern is i<i′ ₁ , then S _i ′ ₁ Same as i′ _l <i
If ＜i _l+1 (1≦l≦M−1), then it is the same as S _i ′ _l , _i ′ _M ＜
i
If so, it is mapped in the same way as Si′M. The stroke correspondence determination result is sent to the stroke feature point resynthesizing device 7.
sent to.

FIG. 2 shows an example of the stroke correspondence between the input pattern "hand" with the incorrect number of strokes written with three strokes and the standard pattern "hand" correctly written with four strokes. Double circles represent correspondence in the first stage, and single circles represent correspondence in the second stage. The numbers attached to each stroke in the figure indicate the stroke order, and the black circles on the strokes indicate the positions of feature points.

The stroke feature point resynthesis device 7 extracts the stroke feature points of the input pattern sent from the stroke feature point extraction device 3 and the standard pattern storage device 4.
The stroke feature points of the standard pattern sent out from the stroke matching device 6 are temporarily stored, and based on the stroke matching result sent out from the stroke matching device 6, one of the opponent patterns among the patterns with a larger number of strokes is selected.
Combine multiple strokes corresponding to one stroke according to the stroke order, convert them into a feature point sequence representing one stroke, and then rearrange the feature point sequence so that the stroke order is the same as the other pattern. device, and when M≧N, q strokes of the standard pattern are added to the i-th stroke Si of the input pattern.
If Rji, R _ij+1 , …, R _ij+q-1 correspond
The stroke feature points of Rji, ..., R _ij+q-1 are connected according to the stroke order and converted into one stroke R'i,
When M<N, the jth stroke of the standard pattern
Rj is the r strokes of the input pattern Sij, S _ij+1 ,
…, S _ij+r-1 if Sij, …, S _ij+r-1
The stroke feature points of are connected in accordance with the stroke order and converted into one stroke S'j. When M≧N, N stroke pairs Si, R′i (1≦i
≦N), but when M<N, M stroke pairs S'j, Rj (1≦j≦M) with the same stroke order are sent to the inter-pattern distance calculation device 8.

The inter-pattern distance calculation device 8 calculates the inter-stroke distance for each corresponding stroke using the input pattern with the same number of strokes and stroke order sent from the stroke feature point resynthesis device 7 and the feature point sequence of the standard pattern. This is a device that calculates and calculates the summation result as the distance between patterns.The distance between strokes is calculated by matching the starting point and ending point, which is a known technique, and applying the DP matching method based on the smaller number of feature points, and calculating the distance between M≧N. When the above Si=
(x′ _1i , y′ _1i , …, x′ _oi , y _oi ) and R′i=(x′ _1i , y
′ _1i ,
..., x′ _si , y ^′ _si ), for example, using the squared distance, if min (n, s) ₌ n, then d ^* _ii = min ^r _o 〓 ^{k= 1} {(x _ki −x′〓 _(k)i ) ² + (y _ki −y′〓 _(k)i ² } (however, τ(1)=1, τ−(n)=s), The inter-pattern distance is calculated as _M 〓 ^j=1 d ^* _ii , and when M < N, the inter-stroke distance d ^* _jj is similarly calculated between S′j and Rj (1≦j≦M). , the distance between the patterns is _M 〓 ^j=1 d ^* _jj .

The calculated inter-pattern distance value is sent to the minimum distance value detection device 9.

The operations of devices 5 to 8 described above are performed for all standard patterns stored in the standard pattern storage device 4. Once this operation is complete,
The minimum distance value detection device 9 detects the minimum distance value among all the accumulated inter-pattern distance values, and outputs the character category of the standard pattern corresponding to this as an identification result.

(5) Description of effects As explained in detail above, according to the present invention,
It is possible to significantly improve the online handwritten character recognition function by allowing writing in the wrong order of strokes and with the wrong number of strokes, which has been considered impossible until now. In particular, it is only necessary to register only the correct stroke order and number of strokes as the standard pattern, and it is necessary to determine a good stroke correspondence between the input pattern of incorrect stroke order and number of strokes and the standard pattern so that there are no omissions or overlaps with each other. , when multiple strokes correspond to one stroke in the opponent's pattern, by combining them into one stroke, the handwriting deformation caused by the wrong number of strokes is absorbed, and the stroke order and number of strokes are aligned. As a result, high recognition accuracy can be achieved, which is a great advantage when applied in the field of online handwritten character recognition, which deals with an unspecified number of scribes who often write with various stroke orders and numbers of strokes.

[Brief explanation of drawings]

FIG. 1 is a functional block diagram of an apparatus used in an embodiment of the present invention, and FIG. 2 is a conceptual diagram of a specific example for explaining the operation of the apparatus shown in FIG. In the figure, 1 is a character information input device, 2 is a preprocessing device, 3 is a stroke feature point extraction device, 4 is a standard pattern storage device, 5 is a distance between strokes calculation device, 6 is a stroke matching device, and 7 is a stroke feature 8 is a point resynthesizing device, 8 is an inter-pattern distance calculating device, and 9 is a minimum distance value detecting device.

Claims (1)

[Claims] 1. It should be recognized that this is a stage in which feature point sequences are extracted and stored at equal intervals with a predetermined increment distance value for all strokes of an average pattern written with the correct number of strokes and stroke order for each character category as a standard pattern. a step of extracting and temporarily storing a sequence of feature points at equal intervals for all strokes of the input pattern using the step distance value; After calculating the distance between strokes including the starting points based on the number of feature points that is smaller in each of the two, the distance between the strokes in the opponent pattern is calculated for all strokes of the pattern with the smaller number of strokes between the input pattern and the standard pattern. The step of selecting the same number of strokes with the minimum sum of inter-stroke distances and having no overlap from each other and associating them with each other, and selecting the stroke order that is closest to the strokes that remained uncorresponded in the above matching among the patterns with a larger number of strokes. The step of assigning the same correspondence as the already corresponding stroke of the preceding pattern, and combining the strokes of the pattern with the larger number of strokes that overlap and correspond to one stroke of the opponent pattern according to the stroke order into one stroke. A step of converting into a feature point sequence representing the stroke of a step of repeatedly performing all the steps described above for all of the standard patterns to be matched and accumulating the addition results sequentially; and a step of calculating the addition results after completing the step; An online handwritten character recognition method comprising the step of: detecting the minimum one among them and outputting a character category of a standard pattern corresponding to the minimum addition result as a recognition result.