ForCenNet: Foreground-Centric Network for Document Image Rectification

TL;DR

ForCenNet introduces a foreground-centric approach to document image rectification, emphasizing foreground elements to improve geometric correction and achieve state-of-the-art results on multiple benchmarks.

Contribution

The paper presents a novel foreground-centric label generation, mask mechanism, and curvature loss to enhance document rectification accuracy.

Findings

Achieves new state-of-the-art on four benchmarks.

Effectively undistorts text lines and table borders.

Outperforms existing methods in geometric correction.

Abstract

Document image rectification aims to eliminate geometric deformation in photographed documents to facilitate text recognition. However, existing methods often neglect the significance of foreground elements, which provide essential geometric references and layout information for document image correction. In this paper, we introduce Foreground-Centric Network (ForCenNet) to eliminate geometric distortions in document images. Specifically, we initially propose a foreground-centric label generation method, which extracts detailed foreground elements from an undistorted image. Then we introduce a foreground-centric mask mechanism to enhance the distinction between readable and background regions. Furthermore, we design a curvature consistency loss to leverage the detailed foreground labels to help the model understand the distorted geometric distribution. Extensive experiments demonstrate that ForCenNet achieves new state-of-the-art on four real-world benchmarks, such as DocUNet, DIR300, WarpDoc, and DocReal. Quantitative analysis shows that the proposed method effectively undistorts layout elements, such as text lines and table borders. The resources for further comparison are provided at https://github.com/caipeng328/ForCenNet.