AERO-HPR

About

Welcome to AERO-HPR: The 1st Workshop on Human Perception and Recognition in Aerial Surveillance at CVPR 2026!

Aerial person recognition has emerged as a critical research area at the intersection of computer vision and biometrics, driven by commercial UAV accessibility and applications ranging from search-and-rescue to public safety. However, state-of-the-art ground-based methods experience substantial performance degradation when applied to aerial data due to extreme viewing angles, atmospheric turbulence, altitude-induced resolution gaps, and motion blur.

This workshop provides the first dedicated forum addressing the complete aerial person analysis pipeline—including detection, tracking, recognition (face, gait, whole-body), re-identification, action recognition, and group analysis—with explicit focus on aerial-specific technical challenges. We bring together researchers from computer vision, biometrics, and surveillance communities to advance novel methods, establish standardized evaluation protocols, and foster responsible development.

The workshop features leading researchers from major initiatives including IARPA BRIAR, DetReIDX, and AG-ReID. Ethics is a central workshop theme, with 25% of program time dedicated to a panel discussion on "Responsible Development of Aerial Person Recognition Systems" addressing privacy protection, algorithmic fairness, and regulatory compliance.

Topics of Interest

Topics include, but are not limited to:

• Aerial Person Detection: Small object detection, multi-scale detection, non-uniform distribution handling, low-resolution challenges
• Aerial Person Tracking: Multi-object tracking, long-term tracking, occlusion handling, trajectory prediction, track association across cameras
• Aerial Face Recognition: Face recognition at altitude, extreme pose variations, low-resolution face matching, atmospheric degradation
• Aerial Gait Recognition: Motion pattern analysis, view-invariant gait features
• Aerial Wholebody Recognition: Full-body biometric recognition, multi-biometric fusion
• Aerial Person Re-identification: Cross-camera matching, aerial-ground re-identification, attribute-based recognition, long-term re-identification
• Aerial Soft Biometric Analysis: Soft biometric attributes such as skin colour, eye colour, hair colour, height, weight, gender, race
• Aerial Human Action Recognition: Activity recognition, gesture recognition, behavior analysis, anomaly detection
• Aerial Crowd Analysis: Crowd counting, group activity recognition, social interaction analysis, crowd behavior modeling
• Aerial Imagery Generation: Image generation, video generation, using synthetic imagery to benefit aerial data analysis
• Multimodal Large Language Models: Training/utilizing LLMs for Aerial Image Analysis
• Ethics and Privacy of Aerial Surveillance: Privacy protection, responsible deployment, algorithmic fairness, regulatory compliance