Wei, Chuheng (University of California, Riverside), Wu, Guoyuan (University of California-Riverside), Barth, Matthew (University of California-Riverside)

RAF-RCNN: Adaptive Feature Transfer from Clear to Rainy Conditions for Improved Object Detection

Scheduled for presentation during the Poster Session "Perception - Road and weather conditions" (WeBT16), Wednesday, September 25, 2024, 14:30−16:30, Foyer

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada

This information is tentative and subject to change. Compiled on October 3, 2024

Abstract

In the challenging realm of object detection under rainy conditions, essential for autonomous driving systems, visual distortions significantly hinder accuracy. This paper introduces Rain Adapt Faster RCNN (RAF-RCNN), an innovative end-to-end approach that effectively merges advanced deraining techniques with robust object detection. Our method streamlines the model architecture by integrating rain removal and object detection into a single process, using a novel feature transfer learning approach for enhanced robustness under rainy conditions. By employing the Extended Area Structural Discrepancy Loss (EASDL), RAF-RCNN enhances feature map evaluation, leading to significant performance improvements. In quantitative testing of the Rainy KITTI dataset, RAF-RCNN achieves a mean Average Precision (mAP) of 51.4% at IOU [0.5, 0.95], exceeding previous methods by at least 5.5%. These results demonstrate RAF-RCNN's potential to significantly enhance perception systems in intelligent transportation, promising substantial improvements in reliability and safety for autonomous vehicles operating in varied weather conditions.