Complex scene character recognition method and system based on class label code representation

Abstract

The invention provides a complex scene character recognition method and system based on class label code representation, and the method comprises the steps: designing a five-stroke class label code representation module, and carrying out the coding of a character according to the designed five-stroke class label code representation module; inputting a scene character image for preprocessing to obtain a training scene character image; inputting the training scene text image into a semantic segmentation network for pixel-level classification to obtain a classification result of the encoded label; learning a transfer matrix of the five-stroke class label codes by utilizing a probability module, and training the whole model to obtain a scene character recognition model; inputting the test image into a trained scene character recognition model for classification to obtain a final semantic segmentation image; carrying out voting on pixels of the semantic segmentation image, and selecting theprediction with the highest vote number as a final character recognition result. According to the technical scheme of the invention, the method is higher in recognition accuracy for the character image of a complex scene, and reduces the complexity of parameters and calculation.

Description

technical field [0001] The invention relates to the technical field of image character recognition, in particular to a method and system for character recognition in complex scenes based on code representation of class marks. Background technique [0002] Natural scene text recognition is an important research content of computer vision, and has a wide range of applications in human-computer interaction, road sign recognition, image retrieval, automatic driving and other fields. Scene images have problems such as complex background, various characters, distortion, occlusion, uneven illumination and low resolution, which makes the recognition of complex scene text a particularly challenging task. [0003] Traditional character recognition technology relies on manual features and specific scenarios, which has great limitations. In recent years, methods based on deep learning have been introduced into the field of scene text detection and recognition and achieved excellent res...

AI Technical Summary

Problems solved by technology

Most of these encoding methods are used in input methods, and the encoding representation of Chi

Images