图文识别方法、装置、计算机设备及存储介质
By preprocessing images from the banking sector and combining an optimized CRNN network with a BERT language model for text recognition, the problem of inaccurate image-text recognition in the banking sector has been solved, achieving accurate recognition results under various interference factors.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- MJOYS COM
- Filing Date
- 2022-08-01
- Publication Date
- 2026-07-17
AI Technical Summary
Existing image recognition methods are inaccurate in the banking sector due to factors such as brightness, color difference, wrinkles, surface camouflage, and similar shapes, making it difficult to accurately recognize text within images in the banking field.
Image preprocessing techniques are used to perform grayscale conversion, size scaling, tilt and rotation correction, and text region localization on images in the banking field. An optimized CRNN network and BERT language model are combined for text recognition. Correction processing is performed through image-text table recognition and field merging to improve recognition accuracy.
It achieves accurate text recognition within images in the banking sector, even under interference from factors such as brightness, color difference, wrinkles, surface masking, and similar shapes, thus improving recognition accuracy.
Smart Images
Figure CN115311666B_ABST