Gait Analysis

Capstone Project on Gait recognition and Analysis for Person Identification at PES University by team PW22_DS_06

Planned Work

• View transformation (This is a major part of the project which would help make it very versatile)

Ongoing Work

• Using Siamese Twins model to measure the effectiveness in similarity prediction between two gait samples

Completed Work

• Data acquisition
• Feature engineering
  • Gait Energy Image
  • Change Energy Image
  • Frame Diffrence Energy Image
• Data preprocessing for model training (Done only for images with a 90 Degree view)
• Testing of Siamese-Twins model on a very small subset of data
• Writing a robust script for training data generation

Technolgies used

• Python 3.6 - 3.9
• tensorflow_gpu-2.6.0
• TensorFlow 2.6.0
• Bazel 3.7.2 (Build Tool)
• MSVC 2019 (Compiler)
• cuDNN 8.1
• CUDA 11.2

Database Links

Anchor Data

Positive

Negative