LRANet: Towards Accurate and Efficient Scene Text Detection with Low-Rank Approximation Network

TL;DR

LRANet introduces a low-rank approximation-based shape modeling and a dual assignment scheme to enhance the accuracy and speed of scene text detection, especially for arbitrary-shaped texts.

Contribution

The paper proposes a novel low-rank approximation method for text shape representation and a dual assignment scheme to improve detection accuracy and inference speed.

Findings

Achieves superior accuracy on challenging benchmarks.

Demonstrates significant speed improvements over existing methods.

Robust in modeling arbitrary-shaped texts.

Abstract

Recently, regression-based methods, which predict parameterized text shapes for text localization, have gained popularity in scene text detection. However, the existing parameterized text shape methods still have limitations in modeling arbitrary-shaped texts due to ignoring the utilization of text-specific shape information. Moreover, the time consumption of the entire pipeline has been largely overlooked, leading to a suboptimal overall inference speed. To address these issues, we first propose a novel parameterized text shape method based on low-rank approximation. Unlike other shape representation methods that employ data-irrelevant parameterization, our approach utilizes singular value decomposition and reconstructs the text shape using a few eigenvectors learned from labeled text contours. By exploring the shape correlation among different text contours, our method achieves consistency, compactness, simplicity, and robustness in shape representation. Next, we propose a dual assignment scheme for speed acceleration. It adopts a sparse assignment branch to accelerate the inference speed, and meanwhile, provides ample supervised signals for training through a dense assignment branch. Building upon these designs, we implement an accurate and efficient arbitrary-shaped text detector named LRANet. Extensive experiments are conducted on several challenging benchmarks, demonstrating the superior accuracy and efficiency of LRANet compared to state-of-the-art methods. Code is available at: \url{https://github.com/ychensu/LRANet.git}