Method for identifying adhered handwritten English characters
A recognition method and a technology of concatenated characters, which are applied in the fields of image processing, text recognition and deep learning, can solve problems such as difficult segmentation of concatenated characters, and achieve the effect of improving the overall recognition rate, ensuring accuracy, and improving the accuracy of segmentation
Pending Publication Date: 2022-01-14
BEIJING UNIV OF TECH
View PDF0 Cites 2 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] Aiming at the difficult segmentation problem existing in the identification of cohesive characters at present, the present invention combines the morphological characteristics of cohesive characters and finds candidate segmentation points based on the structural features of the image, and uses this as the initial segmentation point of the segmentation algorithm to design the segmentation The rules determine the segmentation path. At the same time, for different types of sticky characters, the multi-strategy segmentation method is used for precise segmentation. Finally, the optimal segmentation path is determined by constructing the evaluation method of the segmentation path, and the single character after segmentation is obtained. For the segmented handwritten English characters, the convolutional neural network is used to train the recognition model to recognize individual characters, and a recognition method for sticking handwritten English characters is realized.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0027] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.
[0028] The flow process of the method involved in the present invention comprises the following steps:
[0029] (1) Extract candidate segmentation points based on structural features
[0030] Determine the cohesion point of the cohesive characters and use it as a candidate segmentation point, the specific method is:
[0031] a. Analysis of feature point location area
[0032] In the thinned character image, the change of the pixels in the 8-neighborhood range of the current pixel point can be calculated, and the stroke connection of the point can be analyzed to obtain the feature points in the image. Set the current pixel point as P (x,y) , then x in formula 1 i(i＝1,2…8) is its neighbor point, and P (x,y) =x 9 . where x i = 1 means that the current pixel is the background pixel, x i =0 means black pixels. When the calculated N(P (x,y...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a method for identifying adhered handwritten English characters, which is used for segmenting and identifying the adhered characters. The method comprises the following steps: firstly, finding a candidate segmentation point based on structural features of an image, taking the candidate segmentation point as an initial segmentation point of a segmentation algorithm, and designing a segmentation rule to determine a segmentation path; performing accurate segmentation on different types of adhesion characters by adopting a multi-strategy segmentation mode; and finally, determining an optimal segmentation path by constructing a segmentation path evaluation method; and sending the segmented handwritten English characters into a recognition model obtained by training the convolutional neural network for recognition, and finally obtaining a recognition result of the adhered handwritten English characters.
Description
technical field [0001] The invention relates to the fields of image processing, character recognition and deep learning, and relates to a method for recognizing glued handwritten English characters. Background technique [0002] Handwritten English recognition is an important task in computer vision, especially in automatic review systems. The study found that there are obvious differences in English writing styles at home and abroad. Research on handwriting recognition based on domestic English writing habits is of great significance for improving the recognition rate of handwritten English. During the development of handwriting recognition in recent years, one of the main recognition methods is segmentation first and then recognition, and the other is recognition without segmentation. Due to the blindness and non-standardization of people's writing process, the recognition effect is not ideal. At present, handwritten English recognition is still a big problem in artifici...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06V10/774G06V10/26G06V10/44G06V30/148
CPCG06F18/214
Inventor 付鹏斌宋冬雪杨惠荣
Owner BEIJING UNIV OF TECH