RVAFM: Re-parameterizing Vertical Attention Fusion Module for Handwritten Paragraph Text Recognition

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zheng, Jinhui, Liu, Zhiquan, Si, Yain-Whar, Li, Jianqing, Zhang, Xinyuan, Li, Xiaofan, Huang, Haozhi, Gong, Xueyuan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916644245209088
author Zheng, Jinhui
Liu, Zhiquan
Si, Yain-Whar
Li, Jianqing
Zhang, Xinyuan
Li, Xiaofan
Huang, Haozhi
Gong, Xueyuan
author_facet Zheng, Jinhui
Liu, Zhiquan
Si, Yain-Whar
Li, Jianqing
Zhang, Xinyuan
Li, Xiaofan
Huang, Haozhi
Gong, Xueyuan
contents Handwritten Paragraph Text Recognition (HPTR) is a challenging task in Computer Vision, requiring the transformation of a paragraph text image, rich in handwritten text, into text encoding sequences. One of the most advanced models for this task is Vertical Attention Network (VAN), which utilizes a Vertical Attention Module (VAM) to implicitly segment paragraph text images into text lines, thereby reducing the difficulty of the recognition task. However, from a network structure perspective, VAM is a single-branch module, which is less effective in learning compared to multi-branch modules. In this paper, we propose a new module, named Re-parameterizing Vertical Attention Fusion Module (RVAFM), which incorporates structural re-parameterization techniques. RVAFM decouples the structure of the module during training and inference stages. During training, it uses a multi-branch structure for more effective learning, and during inference, it uses a single-branch structure for faster processing. The features learned by the multi-branch structure are fused into the single-branch structure through a special fusion method named Re-parameterization Fusion (RF) without any loss of information. As a result, we achieve a Character Error Rate (CER) of 4.44% and a Word Error Rate (WER) of 14.37% on the IAM paragraph-level test set. Additionally, the inference speed is slightly faster than VAN.
format Preprint
id arxiv_https___arxiv_org_abs_2503_03104
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RVAFM: Re-parameterizing Vertical Attention Fusion Module for Handwritten Paragraph Text Recognition
Zheng, Jinhui
Liu, Zhiquan
Si, Yain-Whar
Li, Jianqing
Zhang, Xinyuan
Li, Xiaofan
Huang, Haozhi
Gong, Xueyuan
Computer Vision and Pattern Recognition
Artificial Intelligence
Handwritten Paragraph Text Recognition (HPTR) is a challenging task in Computer Vision, requiring the transformation of a paragraph text image, rich in handwritten text, into text encoding sequences. One of the most advanced models for this task is Vertical Attention Network (VAN), which utilizes a Vertical Attention Module (VAM) to implicitly segment paragraph text images into text lines, thereby reducing the difficulty of the recognition task. However, from a network structure perspective, VAM is a single-branch module, which is less effective in learning compared to multi-branch modules. In this paper, we propose a new module, named Re-parameterizing Vertical Attention Fusion Module (RVAFM), which incorporates structural re-parameterization techniques. RVAFM decouples the structure of the module during training and inference stages. During training, it uses a multi-branch structure for more effective learning, and during inference, it uses a single-branch structure for faster processing. The features learned by the multi-branch structure are fused into the single-branch structure through a special fusion method named Re-parameterization Fusion (RF) without any loss of information. As a result, we achieve a Character Error Rate (CER) of 4.44% and a Word Error Rate (WER) of 14.37% on the IAM paragraph-level test set. Additionally, the inference speed is slightly faster than VAN.
title RVAFM: Re-parameterizing Vertical Attention Fusion Module for Handwritten Paragraph Text Recognition
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2503.03104