Poses-based gait recognition

• Grant number: 20/14420-5
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: March 01, 2021
• End date: November 30, 2021
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Aparecido Nilceu Marana
• Grantee: Daniel Ricardo dos Santos Jangua
• Host Institution: Faculdade de Ciências (FC). Universidade Estadual Paulista (UNESP). Campus de Bauru. Bauru , SP, Brazil

Abstract

In recent decades, biometrics has become an important tool for identifying individuals in the most diverse areas, used to prevent fraud and increase the security of citizens in society. Current biometric systems are mostly based on the analysis of physical characteristics such as fingerprints and faces. Despite the good results obtained, resulting from the most recent advances in this area, there is still an important challenge to be overcome: the automatic identification of individuals in low-resolution videos, at a distance, in a covert and non-invasive manner, and without cooperation from individuals being identified. In this scenario, in which conventional biometric systems are not effective, gait recognition is highlighted. The objective of this research project is to propose gait recognition methods based on 2D poses obtained from videos through pose estimation methods, such as OpenPose and PifPaf. The methods will be evaluated on public gait databases, such as CASIA Gait Dataset-A and Dataset-B, and compared with state-of-the-art methods, particularly those based on poses. (AU)