Gait recognition using GEI and pattern trace transform

In this paper, gait recognition using Gait Energy Image (GEI) and pattern trace transform, is presented for human identification. The first step, gait recognition using GEI and pattern trace transform, used silhouette input. The gait period could be estimated by computing the median of the walking-step width of the three consecutive maxima. Subsequently, we calculated the Gait Energy Images from the silhouettes. GEI was transformed to a trace transform image. The pattern trace transform images were calculated using a threshold from the trace transform image to develop the pattern trace transform, which was a significant difference. Finally, we used template matching for identification. The results of gait recognition using GEI and pattern trace transform show that the proposed technique is quite effective and can be developed for higher performance.