Prior-based Domain Adaptive Object Detection for Hazy and Rainy Conditions
This work addresses domain adaptation for object detection in adverse weather, which is crucial for autonomous driving and surveillance, but it is incremental as it builds on existing adversarial and prior-based techniques.
The paper tackles the problem of object detection in hazy and rainy conditions by proposing an unsupervised prior-based domain adversarial framework, achieving improved performance on datasets like Foggy-Cityscapes and RTTS with concrete gains over baseline methods.
Adverse weather conditions such as haze and rain corrupt the quality of captured images, which cause detection networks trained on clean images to perform poorly on these images. To address this issue, we propose an unsupervised prior-based domain adversarial object detection framework for adapting the detectors to hazy and rainy conditions. In particular, we use weather-specific prior knowledge obtained using the principles of image formation to define a novel prior-adversarial loss. The prior-adversarial loss used to train the adaptation process aims to reduce the weather-specific information in the features, thereby mitigating the effects of weather on the detection performance. Additionally, we introduce a set of residual feature recovery blocks in the object detection pipeline to de-distort the feature space, resulting in further improvements. Evaluations performed on various datasets (Foggy-Cityscapes, Rainy-Cityscapes, RTTS and UFDD) for rainy and hazy conditions demonstrates the effectiveness of the proposed approach.