Robust Pedestrian Detection in Inclement Weather and Occluded Scenarios: A Review

Detecting pedestrians is a specialised form of object detection that relies on visual cues to recognize and locate pedestrians. It has practical applications in areas such as artificial intelligence, automated driving, and video surveillance. However, despite notable advancement in the field of deep learning, the task of detecting pedestrians continues to encounter difficulties such as occlusions and poor weather conditions. The objective of this paper is to present a concise overview of recent developments in using deep learning techniques to address these challenges like occlusion and inclement weather in pedestrian detection. Additionally, it presents a compilation of commonly used datasets for detecting pedestrians and an analysis of the miss rates (MR) of chosen approaches utilizing CityPersons dataset, considering different occlusion degrees. Finally, potential scenarios where these strategies could be applied are examined.