Online handwritten Chinese character recognition algorithm and visual key stroke evaluation method

Abstract

The invention relates to an online handwritten Chinese character recognition algorithm and a visual key stroke evaluation method, and belongs to the technical field of mode recognition. According to the local self-attention mechanism provided by the invention, the attention weight corresponding to the state of the hidden layer at each moment can be generated through the neural network, and the recognition precision of online handwritten Chinese characters can be effectively improved through the attention mechanism. The invention provides a visual key stroke evaluation method in online handwritten Chinese character recognition, and the method comprises the steps: displaying strokes (track points or track segments), which play a key role in recognition, in input character sample track coordinates through a weight generated by a self-attention mechanism; key strokes in an online handwritten Chinese character sample can be evaluated more visually, and the mode of learning the character sample through the neural network is analyzed.

Description

technical field [0001] The invention belongs to the technical field of pattern recognition, and in particular relates to an online handwritten Chinese character recognition algorithm and a visual key stroke evaluation method. Background technique [0002] With the rapid development of computer technology, especially smart terminals such as mobile phones play an increasingly important role in people's lives, handwriting input has become an important input method for smart terminals. Specifically, handwritten text recognition refers to the process of allowing the sensor to perceive the writing track of a finger or pen, or obtaining the image of the written text or symbols through a camera device, and the computer recognizes the text through a recognition algorithm. Online handwritten Chinese characters mainly refer to Chinese characters written on the touch screen. The data of such handwritten Chinese characters generally includes time series information such as stroke order a...

AI Technical Summary

Problems solved by technology

[0007] (1) The existing online handwritten Chinese character recognition algorithm only continuously improves the recognition accuracy by improving the algorithm, and does not analyze the key strokes of the character sample trajectory itself

[0008] (2) The existing online handwritten Chinese character recognition algorithm based on recurrent neural network does not introduce a self-attention mechanism to fuse the state vectors of the hidden layer at each moment. By introducing a self-attention mechanism, different trajectory points in the input sample can be distinguished The importance of identifying samples

Images