Recurrent Attention Models for Depth-Based Person Identification
Preprint
Abstract
We present an attention-based model that reasons on human body shape and motion dynamics to identify individuals in the absence of RGB information, hence in the dark. Our approach leverages unique 4D spatio-temporal signatures to address the identification problem across days. Formulated as a reinforcement learning task, our model is based on a combination of convolutional and recurrent neural networks with the goal of identifying small, discriminative regions indicative of human identity. We demonstrate that our model produces state-of-the-art results on several published datasets given only depth images. We further study the robustness of our model towards viewpoint, appearance, and volumetric changes. Finally, we share insights gleaned from interpretable 2D, 3D, and 4D visualizations of our model's spatio-temporal attention.Keywords
All Related Versions
This publication has 59 references indexed in Scilit:
- Deep feature learning with relative distance comparison for person re-identificationPattern Recognition, 2015
- Re-identification with RGB-D Sensors2012
- Robust 3D Action Recognition with Random Occupancy Patterns2012
- Person Identification Using Full-Body Motion and Anthropometric Biometrics from Kinect Videos2012
- Stacked Convolutional Auto-Encoders for Hierarchical Feature Extraction2011
- Viewpoint Invariant Pedestrian Recognition with an Ensemble of Localized Features2008
- Local velocity-adapted motion events for spatio-temporal recognitionComputer Vision and Image Understanding, 2007
- On Space-Time Interest PointsInternational Journal of Computer Vision, 2005
- Gender differences in three dimensional gait analysis data from 98 healthy Korean adultsClinical Biomechanics, 2004
- Gait Analysis for Human Identification2003