Saved in:
Bibliographic Details
Main Authors: Chen, Honghui, Qiu, Yuhang, Wang, Jiabao, Chen, Pingping, Ling, Nam
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2405.09125
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917242595180544
author Chen, Honghui
Qiu, Yuhang
Wang, Jiabao
Chen, Pingping
Ling, Nam
author_facet Chen, Honghui
Qiu, Yuhang
Wang, Jiabao
Chen, Pingping
Ling, Nam
contents Scene Text Recognition (STR) is challenging in extracting effective character representations from visual data when text is unreadable. Permutation language modeling (PLM) is introduced to refine character predictions by jointly capturing contextual and visual information. However, in PLM, the use of random permutations causes training fit oscillation, and the iterative refinement (IR) operation also introduces additional overhead. To address these issues, this paper proposes the Hierarchical Attention autoregressive Model with Adaptive Permutation (HAAP) to enhance position-context-image interaction capability, improving autoregressive LM generalization. First, we propose Implicit Permutation Neurons (IPN) to generate adaptive attention masks that dynamically exploit token dependencies, enhancing the correlation between visual information and context. Adaptive correlation representation helps the model avoid training fit oscillation. Second, the Cross-modal Hierarchical Attention mechanism (CHA) is introduced to capture the dependencies among position queries, contextual semantics and visual information. CHA enables position tokens to aggregate global semantic information, avoiding the need for IR. Extensive experimental results show that the proposed HAAP achieves state-of-the-art (SOTA) performance in terms of accuracy, complexity, and latency on several datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2405_09125
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle HAAP: Vision-context Hierarchical Attention Autoregressive with Adaptive Permutation for Scene Text Recognition
Chen, Honghui
Qiu, Yuhang
Wang, Jiabao
Chen, Pingping
Ling, Nam
Computer Vision and Pattern Recognition
Artificial Intelligence
68T01
I.2.10
Scene Text Recognition (STR) is challenging in extracting effective character representations from visual data when text is unreadable. Permutation language modeling (PLM) is introduced to refine character predictions by jointly capturing contextual and visual information. However, in PLM, the use of random permutations causes training fit oscillation, and the iterative refinement (IR) operation also introduces additional overhead. To address these issues, this paper proposes the Hierarchical Attention autoregressive Model with Adaptive Permutation (HAAP) to enhance position-context-image interaction capability, improving autoregressive LM generalization. First, we propose Implicit Permutation Neurons (IPN) to generate adaptive attention masks that dynamically exploit token dependencies, enhancing the correlation between visual information and context. Adaptive correlation representation helps the model avoid training fit oscillation. Second, the Cross-modal Hierarchical Attention mechanism (CHA) is introduced to capture the dependencies among position queries, contextual semantics and visual information. CHA enables position tokens to aggregate global semantic information, avoiding the need for IR. Extensive experimental results show that the proposed HAAP achieves state-of-the-art (SOTA) performance in terms of accuracy, complexity, and latency on several datasets.
title HAAP: Vision-context Hierarchical Attention Autoregressive with Adaptive Permutation for Scene Text Recognition
topic Computer Vision and Pattern Recognition
Artificial Intelligence
68T01
I.2.10
url https://arxiv.org/abs/2405.09125