Human Posture Recognition Based on DAG-SVMS
2014 ◽
Vol 1042
◽
pp. 117-120
◽
Keyword(s):
A posture feature extraction and recognition method in monitoring environment is proposed in this paper which can recognize human shapes and analyze human postures. First contours of moving objects are extracted from two frames of a consecutive monitoring video. Then feature parameters are calculated from boundary contours to construct feature vector. In order to classify moving object and human and analyze postures, a DAG-SVMS is constructed by training 100 sample images. Results demonstrate the validity of this method.
2013 ◽
Vol 291-294
◽
pp. 2492-2495
Keyword(s):
2014 ◽
Vol 556-562
◽
pp. 2829-2833
◽
2020 ◽
Vol 806
◽
pp. 012035
2021 ◽
Vol 1098
(4)
◽
pp. 042065