Method for detecting and identifying continuous segmented texts in image

Abstract

The invention discloses a method for detecting and identifying continuous segmented texts in an image based on SegLink and Attention-based CRNN fusion processing, and belongs to the technical field ofoptical character recognition, aiming to solve the problems of low text detection accuracy, particularly low inclined text detection accuracy, difficulty in positioning, difficulty in font segmentation and low recognition accuracy in OCR information document digitization. The method includes the steps: establishing a SegLink + CRNN model based on a Tensorflow deep learning framework, detecting text lines in an image through a SegLink network; segmenting the segmented text according to lines; extracting single-line text features through a densely connected convolutional neural network; processing the sequence information of the context in the text by the bidirectional recurrent neural network, and adopting the CTC decoding algorithm to avoid the problem of single word segmentation, and eliminate the influence of the single word segmentation link on the recognition accuracy; and further fusing an Attention mechanism during CTC transcription to improve the recognition accuracy for the text sequence characteristics. The method is applicable to printed form and handwritten form recognition, and can be applied to recognition of multilingual texts such as English and Chinese.

Description

technical field [0001] The invention belongs to the technical field of computer vision, target detection and optical character recognition, relates to text detection and recognition in information documents, and in particular to a continuous text detection and recognition method in images based on fusion processing of SegLink and Attention-based CRNN . Background technique [0002] Text detection and recognition in natural scenes is currently one of the most concerned areas in computer vision, which includes two subtasks: text detection and text recognition. [0003] Most of the current text detection methods are based on a bottom-up process, detecting from low-level features such as simple characters and strokes, and then performing non-text filtering, text line construction, and text line verification. The accuracy of this method largely depends on the results of character detection, and the errors of the detection results will continue to accumulate in the bottom-up proc...

AI Technical Summary

Problems solved by technology

[0006] The present invention proposes a continuous text detection and recognition method in an image based on fusion processing of SegLink and Attention-based CRNN. The convolutional neural network performs feature extraction, and the obtained feature sequence is input into the bidirectional long-term short-term memory cyclic neural network to complete the mapping from the feature sequence to the character sequence, and then the character sequence is CTC transcribed to obtain the final recognition result; thus Solve the problems of text detection in the digitization of OCR information documents, especially the low accuracy of oblique text detection, difficult positioning, difficult font segmentation, and low recognition accuracy

Images