A Digital Recognition Method of Character Wheel Image with Fuzzy State

Abstract

The invention relates to a digital recognition method of a character wheel image with fuzzy state, which specifically comprises the following steps: obtaining a training set of a state extractor and a training set of a classifier; using the training set of the state extractor to train the state extractor , use the training set of the classifier to train the classifier; input the sample to be recognized into the trained state extractor to obtain the state variable, input the state variable into the trained classifier, and output the classification result. The present invention performs effective recognition processing on continuously changing images or images without clear classification boundaries, such as the recognition of character wheel image numbers, and uses the constructed state extractor to extract the characteristic state of the sample to be recognized to obtain the state variable, according to the state variable The degree of membership obtained after the membership function sends the samples to be identified to the classifier corresponding to the degree of membership, so that the samples to be identified can be correctly classified.

Description

Technical field [0001] The invention relates to the technical field of image recognition, in particular to a digital recognition method of a character wheel image with fuzzy states. Background technique [0002] The processing of image information is an important subject of today's technology. In recent years, with the development of deep learning methods, more mature solutions have been given to image recognition problems, and they have achieved landing in scenes such as license plate recognition and text recognition. application. [0003] However, the current academic research is mainly about image recognition problems with clear classification boundaries, and there are often some image classification problems with blurred boundaries in actual scenes. For example, the recognition of digits in the character wheel image is a typical problem of blurred image classification, because the rotation of the character wheel is a continuous process, and there may be complete characters in ...

AI Technical Summary

Problems solved by technology

[0004] In the scene of image recognition, if the data in the fuzzy state is ignored, the image recognition model will not be able to accurately and effectively recognize the image in the fuzzy state, and it will be difficult to popularize and apply.

If the data in the fuzzy state is forcibly divided into categories, it will cause the iteration of the model to fail to converge during the training process.

Images