Shape Matching and Object Recognition Using Shape Contexts
IEEE Transactions on Pattern Analysis and Machine Intelligence
A SIFT Descriptor with Global Context
CVPR '05 Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) - Volume 1 - Volume 01
Motion detection: fast and robust algorithms for embedded systems
ICIP'09 Proceedings of the 16th IEEE international conference on Image processing
BRIEF: binary robust independent elementary features
ECCV'10 Proceedings of the 11th European conference on Computer vision: Part IV
Real-time human pose recognition in parts from single depth images
CVPR '11 Proceedings of the 2011 IEEE Conference on Computer Vision and Pattern Recognition
| Hi-index | 0.00 |
Body pixel classification is a multiclass pixel by pixel image segmentation problem that aims to classify each image pixel to its correspondent human body part. In this article we initially adopted for this problem a Multilayer Perceptron neural network (MLP) classifier using back propagation algorithm to learn network weights and biases. Then confidence intervals based on diffMax criterion are computed in order to make classification more certain. This criterion is computed by the difference between the first and second maximum value of MLP output vector. A 92 % correct classification rate was achieved after applying confidence classification. The classification result will be integrated as an input to a human posture recognition system.