DetReIDX: A Stress-Test Dataset for Real-World UAV-Based Person Recognition

Abstract

Person reidentification (ReID) technology is considered to perform relatively well under controlled, ground-level conditions, but also to break down when deployed in challenging real-world settings. This is due to extreme data variability factors such as resolution, viewpoint changes, scale variations, occlusions, and appearance shifts from clothing/session drifts. Also, the publicly available data sets do not realistically incorporate such kinds and magnitudes of variability, which limits the progress of this technology. This paper introduces DetReIDX, a large-scale aerial-ground person dataset, that was explicitly designed as a stress test to ReID under real-world conditions. DetReIDX is a multi-session set that includes over 18 million bounding boxes from 553 identities, collected in seven university campuses from three continents, with drone altitudes between 5.8 and 120 meters. Singularly, as a key novelty, DetReIDX subjects were recorded in (at least) two sessions on different days, with changes in clothing, daylight and location, making it suitable to actually evaluate long-term person ReID. Further, data were annotated from 16 soft biometric attributes and multitask labels for detection, tracking, ReID, and action recognition. In order to provide empirical objective evidence of DetReIDX usefulness, we considered the specific tasks of human detection, ReID and tracking, and observed that SOTA methods catastrophically degrade performance (up to 80% in detection accuracy and over 70% in Rank-1 ReID) when exposed to DetReIDX’s conditions. The dataset, annotations, and official evaluation protocols are publicly available at https://www.it.ubi.pt/DetReIDX/