Abstract
Multi-orientation scene text detection has received increasing research attention. Previous methods directly predict words or text lines typically using quadrilateral shapes. However, most methods neglect the significance of consistent labeling, which is important for maintaining a stable training process, especially when a large amount of data are involved. Here we solve this problem by proposing a novel method, termed Orderless Box Discretization (OBD), which first discretizes the quadrilateral box into several key edges containing all potential horizontal and vertical positions. To decode accurate vertex positions, a simple yet effective matching procedure is proposed for reconstructing the quadrilateral bounding boxes. Our methods avoids the learning ambiguity issue, which has a significant impact for the learning process. Extensive ablation studies are conducted to quantitatively validate the effectiveness of our proposed method. More importantly, based on OBD, we provide a detailed analysis of the impact of a collection of refinements in the hope to inspire others to build state-of-the-art text detectors. Combining both OBD and these useful refinements, we achieve state-of-the-art performance on various benchmarks including ICDAR 2015, and MLT. Our method also won the first place in the text detection task at the recent ICDAR2019 Robust Reading Challenge on Reading Chinese Text on Signboard (ReCTS), further demonstrating its powerful generalization capability. Code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/1912.09629
