Abstract
Human Action Recognition (HAR) has centered the interest of much research in the last years. Most of this interest has been focused on recognizing motion behavior from a single person (run, jump, walk, hand-wave, etc). However, Human Interaction Recognition (HIR) focus on those cases where several people participate in the scene but the action is carried out only by some of them. The HIR-task is a harder problem than the HAR-task and heretofore very modest scores have been achieved on realistic scenarios. In this paper, we present a new approach to the HIR-task from decoupling the camera and subject motion and using SVM multiple instance learning classifiers. In addition, the pyramidal PaHOF descriptor proposed in the HAR-task context has been adapted. Experimental results on the very challenging TV Human Interactions Dataset [12] are shown, supporting the validity of the proposed approach.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
