Advanced search
Start date
Betweenand


Vision through distortions: Atmospheric Turbulence- and Clothing-invariant long-range recognition

Full text
Author(s):
Bertocco, Gabriel ; Andalo, Fernanda ; Boult, Terrance ; Rocha, Anderson
Total Authors: 4
Document type: Journal article
Source: 2024 IEEE INTERNATIONAL WORKSHOP ON INFORMATION FORENSICS AND SECURITY, WIFS 2024; v. N/A, p. 6-pg., 2024-01-01.
Abstract

Long-range recognition is paramount in security-sensitive settings. It faces the hard task of retrieving images from a high-resolution gallery given a probe image affected by distortions due to atmospheric turbulence and different features, such as clothing. This work proposes a novel atmospheric turbulence- and clothing-invariant whole-body model to address the long-range recognition task. It leverages self-defined proxies across different acquisition ranges, a novel way to create diverse batches based on capturing condition and clothing, and a condition- and clothing-aware loss function. As most whole-body benchmarks have limited ranges, we employ the BRIAR dataset for training and evaluation. It comprises identities captured within 100 to 1,000 meters from the camera in various poses, lighting conditions, and clothing variations. Quantitative and qualitative analysis show our model leads to distortion-invariant discriminative features across different recording capturing ranges. It also obtains competitive performance compared to the state-of-the-art benchmarks Market1501, MSMT17, and DeepChange. (AU)

FAPESP's process: 22/02299-2 - Self-supervised learning for biometrics and beyond
Grantee:Gabriel Capiteli Bertocco
Support Opportunities: Scholarships abroad - Research Internship - Doctorate (Direct)
FAPESP's process: 19/15825-1 - Mining persons, objects and places of interest from heterogeneous data sources
Grantee:Gabriel Capiteli Bertocco
Support Opportunities: Scholarships in Brazil - Doctorate (Direct)
FAPESP's process: 23/12865-8 - Horus: artificial intelligence techniques to detect and forestall synthetic realities
Grantee:Anderson de Rezende Rocha
Support Opportunities: Research Projects - Thematic Grants