sellSeminar sellDFKI sellAugmented Vision sellDeep Learning sellGait Recognition

Advances in Gait Recognition (Seminar)

Shino Sam, Muhammad Saif Ullah Khan

Seminar Report

Abstract

This seminar paper presents a comprehensive survey of gait recognition, a biometric method that identifies individuals based on their walking patterns. Gait recognition has gained significant attention due to its non-invasive nature and applicability in various security, surveillance, and medical scenarios. This survey explores the fundamental principles of gait recognition, including the biomechanics of the human gait cycle and the key features used for recognition. We reviewed various methodologies, including model-free and model-based approaches, and examined their intricacies and performance aspects. The paper also examines publicly available gait datasets, evaluation metrics, and performance benchmarks. Through this survey, we aim to provide a detailed understanding of the state-of-the-art in gait recognition, offering insights into ongoing research and potential areas for future exploration.

Topic

Gait Recognition is the task of identifying a person in a video by their gait/walking style. This has applications in forensic identification and crime prevention. Several benchmark datasets have been curated for this task. Deep learning methods have shown significant success on these benchmarks. In this seminar, we will explore recent advances in this field.

Tasks

1. Identify the datasets available for gait recognition​
2. Read ~3 recent papers which use deep learning for this task​
3. Write a report about the topic, including a brief history, datasets, and recent deep learning works in the domain​

Related Literature

• [1] Gait Recognition in the Wild: A Benchmark
• [2] Gait Recognition in the Wild with Dense 3D Representations and A Benchmark​
• [3] OpenGait: Revisiting Gait Recognition Towards Better Practicality​