Improving Skew Detection and Correction in Different Document Images Using a Deep Learning Approach

During the optical scanning of a document using any of the Optical Character Recognition (OCR) devices, it is difficult to ensure that the scanned document will not get skewed. The high-speed scanning of the skewed document affects the text recognition process. Therefore, it is essential to de-skew the document images to improve the accuracy of the text recognition task. This paper presents a deep learning approach to detect the angle of the skewed scanned document written in different languages. Given the skewed images, we first predict the skewed angle of the document, and later we correct the document images by the angle predicted in reverse order. The angle prediction task is considered to be a classification problem with 360 classes. Each predicted entry in the vector represents the probability of the respective classes with each of value lying between 0 and 1.