Paper
16 December 2022 Research on end-to-end document image character recognition based on differentiable binarization
Yu Li, Tianlong Shao, Hao Wang, Xiaobo Xie, Di Fang
Author Affiliations +
Proceedings Volume 12500, Fifth International Conference on Mechatronics and Computer Technology Engineering (MCTE 2022); 125003V (2022) https://doi.org/10.1117/12.2662638
Event: 5th International Conference on Mechatronics and Computer Technology Engineering (MCTE 2022), 2022, Chongqing, China
Abstract
In view of the poor effect of the current image and character recognition, a research method of end-to-end document image and character recognition based on the differentiable binarization is proposed. Combining the differentiable binarization algorithm to collect and recognize the document image features, an end-to-end document image and character feature extraction algorithm is constructed. Based on the calculation results, the document image is segmented, the regional image features are extracted, and the end-to-end document image and character recognition process is simplified, Finally, the experiment proves that the accuracy of the end-to-end document image and character recognition method based on the differentiable binarization can reach more than 75%, and the overall recognition effect is improved by more than 15% compared with the traditional BFGS method, which meets the research requirements.
© (2022) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Yu Li, Tianlong Shao, Hao Wang, Xiaobo Xie, and Di Fang "Research on end-to-end document image character recognition based on differentiable binarization", Proc. SPIE 12500, Fifth International Conference on Mechatronics and Computer Technology Engineering (MCTE 2022), 125003V (16 December 2022); https://doi.org/10.1117/12.2662638
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Image segmentation

Optical character recognition

Image processing

Feature extraction

Detection and tracking algorithms

Image processing algorithms and systems

Video

RELATED CONTENT

Locally adaptive document skew detection
Proceedings of SPIE (April 03 1997)
Practical automatic Arabic license plate recognition system
Proceedings of SPIE (February 17 2011)
Degraded character recognition based on gradient pattern
Proceedings of SPIE (February 26 2010)

Back to Top