Two-person interaction recognition via spatial multiple instance embedding
Journal of Visual Communication and Image Representation
Academic Press Inc.
63 - 73
MetadataShow full item record
Please cite this item using this persistent URLhttp://hdl.handle.net/11693/20726
Abstract In this work, we look into the problem of recognizing two-person interactions in videos. Our method integrates multiple visual features in a weakly supervised manner by utilizing an embedding-based multiple instance learning framework. In our proposed method, first, several visual features that capture the shape and motion of the interacting people are extracted from each detected person region in a video. Then, two-person visual descriptors are formed. Since the relative spatial locations of interacting people are likely to complement the visual descriptors, we propose to use spatial multiple instance embedding, which implicitly incorporates the distances between people into the multiple instance learning process. Experimental results on two benchmark datasets validate that using two-person visual descriptors together with spatial multiple instance learning offers an effective way for inferring the type of the interaction. © 2015 Elsevier Inc.
Showing items related by title, author, creator and subject.
Farooq J.; Ali, M.B. (IEEE Computer Society, 2014)Hand gesture recognition is a natural and intuitive way to interact with the computer, since interactions with the computer can be increased through multidimensional use of hand gestures as compare to other input methods. ...
Erdem, İ. A.; Erdem, M. E.; Atalay, V.; Çetin, A. E. (SPIE, 2004)It is now possible to design real-time, low-cost computer version systems even in personal computers due to the recent advances in electronics and the computer industry. Due to this reason, it is feasible to develop ...
Kılıboz, N. Ç.; Güdükbay, U. (Academic Press, 2015)We propose an approach to recognize trajectory-based dynamic hand gestures in real time for human-computer interaction (HCI). We also introduce a fast learning mechanism that does not require extensive training data to ...