Image recognition method and device, equipment and medium

Abstract

The invention relates to the field of artificial intelligence, and provides an image recognition method and device, equipment and a medium. Comprising the following steps: accurately determining a threshold value of binarization processing by using a normal distribution algorithm, so that the distinguishing between characters and a background is more obvious, and misjudgment is avoided; segmenting the second sample according to the cutting template, so that more targeted identification is realized; using chinese character strokes as basic elements to extract the feature value of each Chinese character feature, so that not only can the difference of Chinese character structures be reflected, but also the common points of similar characters in structures can be reflected; on the basis of the constructed Chinese character samples, further introducing an uncommon character dictionary and a similar character dictionary further, so that targeted training can be carried out; training to obtain a recognition model so as to generate a recognition result, therefore, image recognition is realized in combination with an artificial intelligence means, the accuracy of image text recognition is improved, and the recognition performance is also improved. In addition, the invention also relates to a block chain technology, and the recognition model can be stored in a block chain node.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to an image recognition method, device, equipment and medium. Background technique [0002] In recent years, as the popularity of artificial intelligence has risen, the field of image recognition has gradually attracted people's attention. Image recognition refers to the automatic recognition of text content on image files through computer technology. In daily life, how to efficiently obtain the text content on a large number of image files, and quickly input the text information contained in the image into the system has become an urgent research topic to be solved. [0003] The current manual method requires manual input of image content, which is inefficient. For a business system with a large number of users, it is obviously not advisable to use manual input and verification after the user inputs the image, especially when the image When the file contai...

AI Technical Summary

Problems solved by technology

[0003] The current manual method requires manual input of image content, which is inefficient. For a business system with a large number of users, it is obviously not advisable to use manual input and verification after the user inputs the image, especially when the image When the file contains a lot of text content information, but the actual system needs to use less information, it will undoubtedly waste time and reduce processing efficiency

[0004] However, the existing image recognition schemes adopt more traditional methods in the early data processing and later feature extraction, without considering the particularity of Chinese characters, so the accuracy rate needs to be improved

Images