Gait analysis
NLPR gait database
|
We recently established a large gait database, called the NLPR database. A digital camera (Panasonic NV-DX100EN) fixed on a tripod is used to capture gait sequences in an outdoor environment, where a single subject moves in the field of view without occlusion. All subjects walk on a straight line path at free cadences and in three different viewing angles with respect to the image plane, namely frontally (90o), laterally (0o) and obliquely (45o). The resulting NLPR gait database includes 20 different subjects and 4 sequences per view per subject. The database thus includes a total of 240 (20×4×3) sequences. These images with 24-bit full color are captured at a rate of 25 frames per second and the original resolution is 352×240. The length of each sequence varies with the pace of the walker, but the average is about 90 frames. Some sample images are shown in the following Figure, where the white line with arrow represents the walking path. Apply to the NLPR gait database. |
a.Lateral view b.Oblique view c.Frontal view
Statistical Gait Recognition
|
1. A simple but efficient model-free gait recognition algorithm based on Principal Component Analysis (PCA) is proposed. For each image sequence, an improved background subtraction procedure is first used to accurately extract the spatial silhouettes of a walking figure from the background.Then, eigenspace transformation to time-varying distance signals derived from moving silhouette shapes are performed to realize gait signature extraction. Gait recognition using spatio-temporal correlation matching or the k-nearest neighbor classifier is finally accomplished in the lower-dimensional eigenspace, and some additional personalized physical properties in response to pace, stride and build are selected for the validation of final decision. Experimental results show that the proposed algorithm has an encouraging recognition performance with relatively lower computational cost. 2. A novel gait recognition algorithm based on statistical shape analysis is proposed. For each gait sequence, a background subtraction procedure is used to segment spatial silhouettes of the walking figures from the background. Static pose changes of these silhouettes over time are represented as a sequence of associated complex configurations in a common coordinate, and are then analyzed using the Procrustes shape analysis method to obtain mean shape as gait signature. The k-nearest neighbor classifier and the nearest exemplar classifier based on the full Procrustes distance measure are adopted for recognition. Experimental results demonstrate that the proposed algorithm has an encouraging recognition performance. |
Fusion of Static and Dynamic Body Biometrics for Personal Recognition
|
We propose a visual recognition algorithm using static and dynamic information of body biometrics obtained from walking video. For each sequence involving a walking figure, static pose changes of the segmented moving silhouettes from the background are represented as an associated sequence of complex vector configurations, and are then analyzed using the Procrustes shape analysis method to obtain a compact appearance representation, called static info of body. Also, a model-based approach is presented under a CONDENSATION framework to track the walker and recover joint-angle trajectories of lower limbs, called dynamic information of gait. Both static and dynamic cues are respectively used for personal recognition using the nearest exemplar classifier. They are also effectively fused using different combination rules on decision level to improve the performance of both identification and verification. Experimental results on a dataset including 20 subjects demonstrate the validity of the proposed algorithm. |
Publications
1. L.
Wang, T. Tan, H. Ning and W. Hu, Silhouette Analysis Based Gait
Recognition for Human Identification, IEEE Trans. on Pattern Analysis and
Machine Intelligence, 25(12), 2003. [PDF]
2. L. Wang, H. Ning, T. Tan and W. Hu, Vision-based gait recognition for human identification at a distance, IEEE Trans. Circuits and Systems for Video Technology Special Issue on Image- and Video-Based Biometrics, 14(2), 149-158, 2004.
3. L. Wang, T. N. Tan, W. M. Hu and H. Z. Ning, Automatic Gait Recognition Based on Statistical Shape Analysis, IEEE Trans. on Image Processing, 12(9), 2003. [PDF]
4. L. Wang, H. Ning, T. Tan and W. Hu, Fusion of static and dynamic body biometrics for gait recognition, Intl. Conf. on Computer Vision (ICCV), France, 2003. [PDF]
5.
IEEE Int. Conf. on Image Processing (ICIP), 2002. [PDF]