Multi-language scene character recognition method and recognition system

Abstract

The invention relates to a multi-language scene character recognition method and recognition system. The method comprises the steps that the language type of characters in a scene character image is determined; a deep convolutional neural network model is determined according to the language type of the characters; the convolutional layer characteristics of the scene character image are extracted by using the deep convolutional neural network model; a space pyramid model is established based on the convolutional layer characteristics; high-order coding is performed on each space area on the space pyramid model by using a Gaussian model; the results of high-order coding are spliced to act as scene character descriptors; and the scene character descriptors are classified by using a classifier so as to realize multi-language scene character recognition. The multi-language scene character recognition method and recognition system have great recognition effect on the multi-language scene character image so that the method is a general character recognition method and has great adaptability for multi-language scene character recognition.

Description

technical field [0001] The invention relates to the technical field of text image recognition in the field of pattern recognition and machine vision, in particular to a multilingual scene character recognition method and a recognition system. Background technique [0002] As a symbol designed by humans, text expresses high-level information and is of great help to understanding the content in images. Therefore, the automatic detection and recognition of characters has aroused people's high attention. It has extremely important significance for the storage, classification, understanding and retrieval of images and videos, and has broad application prospects and commercial value. [0003] However, because the characters in natural scenes are affected by factors such as complex background, uneven illumination, low resolution, various fonts, and different colors, the recognition effect is not satisfactory by directly using traditional optical character recognition (OCR) technolo...

AI Technical Summary

Problems solved by technology

These features rely on the BoW framework, and character recognition performance is unsatisfactory due to unsatisfactory lexicon learning and encoding strategies

The other is the feature based on deep learning, which relies on the neural network to extract features from the fully connected layer, so the time complexity is high and the fully connected layer features ignore the spatial structure information of the character, and the spatial information is very important for character recognition. important

Images