Improved Techniques for Learning to Dehaze and Beyond: A Collective Study
This work addresses image restoration and visual understanding in hazy conditions, but it is incremental as it builds on existing datasets and methods.
The paper tackled single image dehazing and object detection in hazy images using the RESIDE dataset, showing that perception-driven loss improves dehazing and domain-adaptive detectors enhance detection, with significant performance gains reported.
Here we explore two related but important tasks based on the recently released REalistic Single Image DEhazing (RESIDE) benchmark dataset: (i) single image dehazing as a low-level image restoration problem; and (ii) high-level visual understanding (e.g., object detection) of hazy images. For the first task, we investigated a variety of loss functions and show that perception-driven loss significantly improves dehazing performance. In the second task, we provide multiple solutions including using advanced modules in the dehazing-detection cascade and domain-adaptive object detectors. In both tasks, our proposed solutions significantly improve performance. GitHub repository URL is: https://github.com/guanlongzhao/dehaze