Poster - Deep Gait Recognition via Millimeter Wave.

The key to personalizable behavior in smart spaces is knowing where and who a particular person is. However, concerns arise around potential leakage of face/video information, and many people do not accept cameras in their homes or workplaces. With the aid of a deep recurrent network, we propose a human recognition system that identifies gaits based on millimeter wave (MMwave). By a commercial, off-the-shelf radar, our system first obtains sparse point clouds from the reflection profiles of people walking. A deep neural network is then used to extract gait information from sequential point clouds and identify different people. Preliminary results demonstrate that MMwave is a very promising modality for gait recognition.