In order to compare and evaluate various gait recognition algorithms, a Human ID gait challenge problem
was designed. It included a carefully collected dataset of a number of people walking on different surface,
using different shoes, at different times and sometimes carrying a load. Baseline performance algorithms and
their results were also provide. Details on the challenge problem can be found on the
Gait Challenge Problem Website .

Below, we present the results on our evaluation algorithms on this dataset.

Results Using Hidden Markov Model:

Results Using Synchronized Matching (Dynamic Time Warping):