The 5^th UG²+ Workshop and Challenge took place at CVPR 2022 Conference, at the New Orleans Ernest N. Morial Convention Center. After two years of the pandemic, CVPR returned back to in-person mode, and thousands of audience attended during the workshop days. With the strong support of NVIDIA, the 5th UG2 Workshop and Challenge set a new record for both the number of attendance and challenge submissions, leading to a successful promotion of GPU-accelerated Deep Learning Models in enhancing degraded image/video processing and understanding.

Scene understanding for applications such as intelligence, surveillance, and reconnaissance (ISR) and autonomous vehicles becomes extremely challenging in adverse weather conditions such as haze, fog and mist. These atmospheric phenomena with smoke particles and minute water droplets often result in imagery with non-linear noise, blur, reduced contrast levels and color dimming issues. Thus, the on-board vision systems are significantly obscured. These visual artifacts generated from uncontrolled and potentially changing environment or Degraded Visual Environments (DVE) poses major challenges in image enhancement and restoration, and object detection and classification, some of the key tasks towards the final goal of semantic scene understanding.

Our UG²+ challenge "Object Detection in Haze" aims to evaluate and advance object detection algorithms’ robustness on images captured from hazy environmental situations. Participants are allowed to use a restoration/enhancement pre-processing step in the detection pipeline. In other words, they will not be tasked with the creation of novel object detection algorithms. A list of detection algorithms will be provided to them in order to facilitate studies of the interaction between image restoration and enhancement algorithms and the detectors. During the evaluation, the selected detection algorithms will be run on the test images. Through this challenge and benchmark, we aim to encourage more state-of-the-art single-image dehazing, haze quantification, and object detection algorithms.

Top teams on the leaderboard are required to submit their code for the validation process. For a fair comparison, we asked the top 15 teams to submit their code. This evaluation process requires heavy computation costs. With NVIDIA’s strong GPU support, we smoothly completed this evaluation and announced the winner and two runner-ups. They all achieved strong performance. Congratulations!

On the CVPR workshop day, hundreds of audiences attended the UG2 workshop. We invited in total of nine speakers, who are all experts in vision and machine learning domains, such as DeRF, Derain, Video Enhancement, Turbulence, ViT, Domain Shift.

Finally, the organizers of the UG2 workshop again sincerely appreciate the strong support from NVIDIA!