Research on Intelligent System of Malformed Document Correction Based on Computer Big Data Technology

The document image captured by the camera may be tilted or deformed, so that the optical character recognition (OCR) software cannot recognize it correctly. In this paper, image segmentation technology is firstly used to detect words and text lines. We then use linear fitting to obtain lower and upper baselines of words. Finally, according to the corrected baseline and vertical displacement distance, the word is rotated and displaced to obtain the corrected image. The experimental simulation shows that the corrected baseline and vertical displacement distance obtained by this method are not affected by the specific text content of the document. In this way, the recognition rate of the corrected image in the optical character recognition stage is greatly improved.