 |
    |
Recognizing Complex Events in Videos by Learning Key Static-Dynamic Evidences
Kuan-Ting Lai1,2, Dong Liu3, Ming-Syan Chen1,2, and Shih-Fu Chang3
1Graduate Institute of Electrical Engineering, National Taiwan University, Taiwan
ktlai@arbor.ee.ntu.edu.tw
mschen@arbor.ee.ntu.edu.tw
2Research Center for IT Innovation, Academia Sinica, Taiwan
3Department of Electrical Engineering, Columbia University, USA
dongliu@ee.columbia.edu
sfchang@ee.columbia.edu
Abstract. Complex events consist of various human interactions with different objects in diverse environments. The evidences needed to recognize events may occur in short time periods with variable lengths and can happen anywhere in a video. This fact prevents conventional machine learning algorithms from effectively recognizing the events. In this paper, we propose a novel method that can automatically identify the key evidences in videos for detecting complex events. Both static instances (objects) and dynamic instances (actions) are considered by sampling frames and temporal segments respectively. To compare the characteristic power of heterogeneous instances, we embed static and dynamic instances into a multiple instance learning framework via instance similarity measures, and cast the problem as an Evidence Selective Ranking (ESR) process. We impose 
1 norm to select key evidences while using the Infinite Push Loss Function to enforce positive videos to have higher detection scores than negative videos. The Alternating Direction Method of Multipliers (ADMM) algorithm is used to solve the optimization problem. Experiments on large-scale video datasets show that our method can improve the detection accuracy while providing the unique capability in discovering key evidences of each complex event.
Keywords: Video Event Detection, Infinite Push, Key Evidence Selection, ADMM
LNCS 8691, p. 675 ff.
lncs@springer.com
© Springer International Publishing Switzerland 2014