《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (3): 933-939.DOI: 10.11772/j.issn.1001-9081.2025040395

• 多媒体计算与计算机仿真 • 上一篇    下一篇

基于组件协同优化剪枝的Transformer图像去雾

郭纪新, 张婷()   

  1. 北京理工大学 自动化学院,北京 100081
  • 收稿日期:2025-04-14 修回日期:2025-06-05 接受日期:2025-06-11 发布日期:2025-06-23 出版日期:2026-03-10
  • 通讯作者: 张婷
  • 作者简介:郭纪新(2000—),男,河南许昌人,硕士研究生,主要研究方向:图像处理、图像去雾
  • 基金资助:
    教育部产学合作协同育人项目(202102023015);北京市高等教育学会面上课题(MS2022045);北京理工大学研究生教研教改面上项目(2025YBJG006)

Transformer image dehazing based on component collaborative optimization pruning

Jixin GUO, Ting ZHANG()   

  1. School of Automation,Beijing Institute of Technology,Beijing 100081,China
  • Received:2025-04-14 Revised:2025-06-05 Accepted:2025-06-11 Online:2025-06-23 Published:2026-03-10
  • Contact: Ting ZHANG
  • About author:GUO Jixin, born in 2000, M. S. candidate. His research interests include image processing, image dehazing.
  • Supported by:
    University-Industry Collaborative Education Program of the Ministry of Education(202102023015);General Program of Beijing Association of Higher Education(MS2022045);General Program of Graduate Teaching and Research Program of Beijing Institute of Technology(2025YBJG006)

摘要:

基于Transformer的图像去雾算法取得了较好去雾效果,但存在网络参数量大和去雾速度慢的问题。为实现对去雾网络冗余部分的定向修剪,在不影响去雾质量的前提下缩短去雾时间,提出一种基于组件协同优化剪枝的Transformer图像去雾方法CCOP-IDT(Component Collaborative Optimization Pruning Image Dehazing Transformer)。首先,采用5级Transformer构建去雾网络预训练模型;其次,将网络剪枝建模为优化问题,使用费雪信息评估权重参数重要度,并利用黑塞矩阵衡量剪枝组件对网络输出的联合影响,从而建立多种剪枝组件的协同优化方法;最后,采用进化算法求解最优剪枝率序列,从而得到预训练模型的最优子网络。实验结果表明,剪枝后的网络参数量控制在0.476×106,相较于剪枝前减少了28.8%,去雾时间缩短了25.0%;在合成有雾数据集RESIDE-6K上,所提方法的峰值信噪比(PSNR)达到29.60 dB,结构相似度(SSIM)达到0.968 7,与剪枝前相比仅分别降低了1.63%和0.46%。可见,在定量和定性评估上,所提方法都表现良好,能够在基本保持量化指标和主观观感的前提下,大幅减少模型参数量,提高图像去雾速度。

关键词: 图像去雾, Transformer, 模型剪枝, 费雪信息, 黑塞矩阵, 进化算法

Abstract:

Image dehazing algorithms based on Transformer achieve good dehazing effects, but there are problems such as large number of network parameters and low dehazing speed. In order to prune redundant parts of the dehazing network directionally and shorten dehazing time without affecting dehazing quality, a Transformer image dehazing method based on component collaborative optimization pruning, CCOP-IDT (Component Collaborative Optimization Pruning Image Dehazing Transformer), was proposed. Firstly, a 5-level Transformer was used to construct a pre-training model of dehazing network. Then, the network pruning was modeled as an optimization problem, Fisher information was used to evaluate the importance of weight parameters, and Hessian matrix was used to measure the joint influence of pruning components on network output, so as to establish a collaborative optimization method for multiple pruning components. Finally, an evolutionary algorithm was employed to solve the optimal pruning rate sequence, so as to obtain the optimal sub-network of the pre-trained model. Experimental results show that after pruning, the number of network parameters is controlled to 0.476×106, which is reduced by 28.8% compared with that before pruning, and the dehazing time is shortened by 25.0%. On the synthetic hazy dataset RESIDE-6K, the proposed method has the Peak Signal-to-Noise Ratio (PSNR) reached 29.60 dB, and the Structural SIMilarity (SSIM) reached 0.968 7, which are reduced by only 1.63% and 0.46% compared with those before pruning, respectively. It can be seen that in terms of both quantitative and qualitative evaluation, the proposed method performs well with great reduction of the model parameters and improvement of the image dehazing speed while maintaining the quantitative indices and subjective perception basically.

Key words: image dehazing, Transformer, model pruning, Fisher information, Hessian matrix, evolutionary algorithm

中图分类号: