Rethinking Appearance-Based Deep Gait Recognition: Reviews, Analysis, and Insights From Gait Recognition Evolution

Area of Science:

• Biometrics and Pattern Recognition
• Computer Vision
• Artificial Intelligence

Background:

• Gait recognition is a key biometric for public security.
• Appearance-based methods using silhouettes are popular and outperform model-based methods.
• Multiframe silhouettes enhance spatiotemporal representation in appearance-based gait recognition.

Purpose of the Study:

• To provide a comprehensive review of multiframe appearance-based gait recognition methods.
• To unify performant models within a single framework.
• To investigate data arrangement effects and the scaling ability of existing methods.

Main Methods:

• Literature review tracing the evolution of gait recognition techniques.
• Development of a unified framework for multiframe appearance-based models.
• Empirical studies on data arrangement and model scalability.

Main Results:

• Identification of key advancements in multiframe appearance-based gait recognition.
• Demonstration of the impact of data arrangement on performance.
• Analysis of the scaling capabilities of current methods.

Conclusions:

• Multiframe appearance-based gait recognition shows significant advancements.
• Further research is needed on data arrangement and scalability.
• The study identifies current challenges and future research directions in gait recognition.