Efficient Human Pose Estimation from Single Depth Images
Jamie Shotton, Ross Girshick, Andrew W Fitzgibbon et al. · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2012
We describe two new approaches to human pose estimation. Both can quickly and accurately predict the 3D positions of body joints from a single depth image without using any temporal information. The key to both approaches is the use of a large, realistic, and highly varied synthetic set of training images. This allows us to learn models that are largely invariant to factors such as pose, body shape, field-of-view cropping, and clothing. Our first approach employs an intermediate body parts representation, designed so that an accurate per-pixel classification of the parts will localize the joints of the body. The second approach instead directly regresses the positions of body joints. By using simple depth pixel comparison features and parallelizable decision forests, both approaches can run super-real time on consumer hardware. Our evaluation investigates many aspects of our methods, and compares the approaches to each other and to the state of the art. Results on silhouettes suggest broader applicability to other imaging modalities.
Cited by 576 works. doi:10.1109/tpami.2012.241
Authors
- Jamie Shotton
- Ross Girshick
- Andrew W Fitzgibbon
- Toby Sharp
- Mat Cook
- Mark Finocchio
- Pushmeet Kohli
- Antonio Criminisi
- Alex Kipman
- Andrew Blake
References
- Random Forests
- Induction of Decision Trees
- The Elements of Statistical Learning: Data Mining, Inference, and Prediction
- Object class recognition by unsupervised scale-invariant learning
- Mean shift: a robust approach toward feature space analysis
- The Elements of Statistical Learning
- Shape matching and object recognition using shape contexts
- The Representation and Matching of Pictorial Structures
- Shape Quantization and Recognition with Randomized Trees
- Pictorial Structures for Object Recognition