Image similarity determination and model training method and device, equipment and medium

Abstract

The invention relates to the technical field of image processing, and discloses an image similarity determination and model training method and device, equipment and a medium. The method comprises the following steps: collecting a plurality of handwritten character images, and performing data augmentation operation based on each handwritten character image and a preset data augmentation model to generate a training sample set, the handwritten character images including handwritten character images easy to recognize and handwritten character images difficult to recognize, and the preset data augmentation model being constructed based on writing kinematics; and on the basis of the training sample set, training a preset neural network model to generate an image similarity determination model. By means of the technical scheme, data augmentation is conducted on a small number of collected handwritten character images based on the writing kinematics principle so as to obtain a large number of training samples, and the model training efficiency and the recognition accuracy of the model for the handwritten character images are improved.

Description

technical field [0001] The present disclosure relates to the technical field of image processing, and in particular to a method, device, device and medium for determining image similarity and model training. Background technique [0002] Character image similarity evaluation is a specific problem of image similarity evaluation. The so-called character image means that the contents of the image are characters, letters or numbers. Character image similarity evaluation has very important value in many problems, such as handwriting comparison, photo search, electronic correction of homework, etc. The handwritten characters in the image can be recognized through character image similarity evaluation, and then the recognized handwritten characters can be compared with their possible standard character images to obtain the similarity between the two, which can be used for subsequent operations such as handwriting comparison, judgment and search. Provides powerful prior knowledge. ...

AI Technical Summary

Problems solved by technology

[0004] However, considering the gap between the writer's handwriting and the standard font, as well as the fact that handwritten characters are extremely difficult to recognize or cannot be recognized due to random operations such as scribbling, sketching, and smearing, the characters of the neural network model based on deep learning Very low confidence in recognition

In order to improve the accuracy of model recognition for difficult-to-recognize handwritten characters, it is necessary to use a large number of training samples for model training. However, the collection of training samples for difficult-to-recognize handwritten characters is difficult and labor-intensive, making model training inefficient

Images