Abstract:More and more end-to-end text spotting methods based on Transformer architecture have demonstrated superior performance. These methods utilize a bipartite graph matching algorithm to perform one-to-one optimal matching between predicted objects and actual objects. However, the instability of bipartite graph matching can lead to inconsistent optimization targets, thereby affecting the training performance of the model. Existing literature applies denoising training to solve the problem of bipartite graph matching instability in object detection tasks. Unfortunately, this denoising training method cannot be directly applied to text spotting tasks, as these tasks need to perform irregular shape detection tasks and more complex text recognition tasks than classification. To address this issue, we propose a novel denoising training method (DNTextSpotter) for arbitrary-shaped text spotting. Specifically, we decompose the queries of the denoising part into noised positional queries and noised content queries. We use the four Bezier control points of the Bezier center curve to generate the noised positional queries. For the noised content queries, considering that the output of the text in a fixed positional order is not conducive to aligning position with content, we employ a masked character sliding method to initialize noised content queries, thereby assisting in the alignment of text content and position. To improve the model's perception of the background, we further utilize an additional loss function for background characters classification in the denoising training <a class="link-external link-http" href="http://part.Although" rel="external noopener nofollow">this http URL</a> DNTextSpotter is conceptually simple, it outperforms the state-of-the-art methods on four benchmarks (Total-Text, SCUT-CTW1500, ICDAR15, and Inverse-Text), especially yielding an improvement of 11.3% against the best approach in Inverse-Text dataset.

Page-Level Handwritten Word Spotting via Discriminative Feature Learning.

Bootstrapping Weakly Supervised Segmentation-free Word Spotting through HMM-based Alignment

HWNet v2: An Efficient Word Image Representation for Handwritten Documents

Handwritten-word spotting using biologically inspired features

Offline Detection of Misspelled Handwritten Words by Convolving Recognition Model Features with Text Labels

Robust End-to-End Offline Chinese Handwriting Text Page Spotter with Text Kernel

Word length-aware text spotting: Enhancing detection and recognition in dense text image

Towards End-to-End Text Spotting in Natural Scenes

Hierarchical Text Spotter for Joint Text Spotting and Layout Analysis

Mask TextSpotter: An End-to-End Trainable Neural Network for Spotting Text with Arbitrary Shapes

A Multi-oriented Chinese Keyword Spotter Guided by Text Line Detection

An End-to-End TextSpotter with Explicit Alignment and Attention.

PageNet: Towards End-to-End Weakly Supervised Page-Level Handwritten Chinese Text Recognition

WaterSeeker: Pioneering Efficient Detection of Watermarked Segments in Large Documents

Improving Image-Text Matching by Integrating Word Sense Disambiguation

Unsupervised Deep Learning for Handwritten Page Segmentation

Exploiting Multi-Scale Fusion, Spatial Attention and Patch Interaction Techniques for Text-Independent Writer Identification

DNTextSpotter: Arbitrary-Shaped Scene Text Spotting via Improved Denoising Training

Weighted-Gradient Features For Handwritten Line Segmentation

Siamese based Neural Network for Offline Writer Identification on word level data

Neural Chinese Word Segmentation with Dictionary Knowledge