Revisiting Document Image Dewarping by Grid Regularization

TL;DR

This paper proposes a novel grid regularization approach for document image dewarping that emphasizes readability by preserving text lines and boundaries through constrained optimization, outperforming prior methods.

Contribution

It introduces a new grid regularization scheme that leverages boundary and text line information for improved dewarping without relying solely on neural network optical flow estimation.

Findings

Outperforms prior arts in readability metrics

Maintains high image quality on DocUNet benchmark

Effective in preserving document structure during dewarping

Abstract

This paper addresses the problem of document image dewarping, which aims at eliminating the geometric distortion in document images for document digitization. Instead of designing a better neural network to approximate the optical flow fields between the inputs and outputs, we pursue the best readability by taking the text lines and the document boundaries into account from a constrained optimization perspective. Specifically, our proposed method first learns the boundary points and the pixels in the text lines and then follows the most simple observation that the boundaries and text lines in both horizontal and vertical directions should be kept after dewarping to introduce a novel grid regularization scheme. To obtain the final forward mapping for dewarping, we solve an optimization problem with our proposed grid regularization. The experiments comprehensively demonstrate that our proposed approach outperforms the prior arts by large margins in terms of readability (with the metrics of Character Errors Rate and the Edit Distance) while maintaining the best image quality on the publicly-available DocUNet benchmark.