arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2401.05907cs.CV
收藏

基于扩散模型的高效图像去模糊网络

Efficient Image Deblurring Networks based on Diffusion Models

College of Information and Electrical Engineering(信息与电气工程学院);China Agricultural University(中国农业大学)

Kang Chen, Yuanjie Liu

AI总结:

本文提出基于扩散模型和滑动窗口Transformer的Swintormer去模糊网络,大幅降低计算与内存需求,提升PSNR,适用于内存受限设备。

摘要:

本文提出了一种用于散焦去模糊的滑动窗口模型,命名为Swintormer,该模型以极低的内存使用量实现了迄今为止最佳的性能。该方法利用扩散模型生成潜在先验特征,有助于恢复更多细节的图像。此外,通过调整滑动窗口策略,它引入了专门的Transformer块以提高推理效率。采用这种新方法使得每次迭代的乘加运算(MACs)大幅减少,显著降低了内存需求。与当前领先的GRL方法相比,我们的Swintormer模型显著减少了必须依赖内存容量的计算负载,从140.35 GMACs降至8.02 GMACs,同时将散焦去模糊的峰值信噪比(PSNR)从27.04 dB提升至27.07 dB。这项创新技术使得在内存受限的设备上处理更高分辨率的图像成为可能,极大地拓宽了潜在的应用场景。文章最后进行了消融研究,全面考察了每个网络模块对最终性能的贡献。源代码和模型将在以下网站提供:https://github.com/bnm6900030/swintormer。

英文摘要:

This article presents a sliding window model for defocus deblurring, named Swintormer, which achieves the best performance to date with remarkably low memory usage. This method utilizes a diffusion model to generate latent prior features, aiding in the restoration of more detailed images. Additionally, by adapting the sliding window strategy, it incorporates specialized Transformer blocks to enhance inference efficiency. The adoption of this new approach has led to a substantial reduction in Multiply-Accumulate Operations (MACs) per iteration, drastically cutting down memory requirements. In comparison to the currently leading GRL method, our Swintormer model significantly reduces the computational load that must depend on memory capacity, from 140.35 GMACs to 8.02 GMACs, while improving the Peak Signal-to-Noise Ratio (PSNR) for defocus deblurring from 27.04 dB to 27.07 dB. This innovative technique enables the processing of higher resolution images on memory-limited devices, vastly broadening potential application scenarios. The article wraps up with an ablation study, offering a comprehensive examination of how each network module contributes to the final performance.The source code and model will be available at the following website: https://github.com/bnm6900030/swintormer.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑