Htike, Kyaw Kyaw and Khalifa, Othman Omran and Lai, Weng Kin
(2011)
Human posture recognition results using database A.
In:
Human Behaviour Recognition, Identification and Computer Interaction.
IIUM Press, Kuala Lumpur, pp. 49-57.
ISBN 978-967-418-156-7
Abstract
Human Posture Recognition gives machines the ability to detect, track and identify people and their actions from video, has become a central topic in computer vision research. Recognition of human posture is a very challenging problem. The training and evaluation
stage, datasets of pre-processed posture images are needed. After the video sequences have been prepared, for the training and evaluation stage, they are pre-processed to produce three different types of datasets. This chapter will explain the results obtained using Dataset A.
This dataset was obtained after pre-processing some of the video sequences that were taken at MIMOS building. In this dataset, there are six types of postures considered. The "unknown" posture was added so that during the training stage, the system could learn to differentiate
postures which cannot be classified as belonging to any of the other types. This dataset is the main dataset for evaluating and comparing the performance (in terms of accuracy, i.e. recognition rate) of the different classifiers
Actions (login required)
 |
View Item |
![[img]](http://irep.iium.edu.my/style/images/fileicons/application_pdf.png)
Download Statistics
Download Statistics