Paper Image

Dual-path deraining via spatial and frequency analysis

Published on:

7 February 2024

Primary Category:

Computer Vision and Pattern Recognition

Paper Authors:

Yuhong He,

Aiwen Jiang,

Lingfang Jiang,

Zhifeng Wang,

Lu Wang

Bullets

Key Details

Proposes network integrating spatial and frequency analysis for image deraining

Introduces new spatial and frequency feature extraction blocks

Uses adaptive fusion module to couple dual-domain information

Achieves top results on public rainy image benchmarks

Shows excellent robustness on downstream vision tasks

AI generated summary

Dual-path deraining via spatial and frequency analysis

This paper develops an effective image deraining network called DPCNet that utilizes both spatial and frequency domain information to remove rain from images. It introduces new modules to extract features from both domains and fuse them adaptively. Experiments show state-of-the-art deraining performance and excellent robustness for downstream computer vision tasks.

Answers from this paper

Comments

No comments yet, be the first to start the conversation...

Sign up to comment on this paper

Sign Up