Fast Action Detection with One Query Example Based on Hough Voting
Detect action in the target video based on a query video is an important research topic. We propose a fast action detection method. First, Features extracted at the interest points from the query video. Then, the clips are formed by sliding a window on the video. For each clip, the points of all the frames are compared with that in the first frame. The matched pairs are counted in the displacement cells to form a displacement histogram. This histogram sequence represents the query video. Then, we divide the target video into cubes. These cubes are similarly represented by histogram sequences. Matrix Cosine Similarity (MCS) is used to compute the similarities between the query video and cubes. Last, we localize the action using the locations of the matched points. Our key contribution is the proposed fast action representation method. Experiments on challenging datasets confirm the effectiveness and efficiency of our method.
KeywordsAction Detection Hough Voting Displacement Histogram Matrix Cosine Similarity
Unable to display preview. Download preview PDF.
- 1.Lampert, C.H., Blaschko, M.B., Hofmann, T.: Beyond Sliding Windows: Object Localization by Efficient Subwindow Search. In: Proc. IEEE Conf. on Computer Vision and Pattern Recognition (2008)Google Scholar
- 2.Gall, J., Lempitsky, V.: Class-specific Hough forests for object detection. In: Proc. IEEE Conf. on Computer Vision and Pattern Recognition (2009)Google Scholar
- 3.Yu, G., Yuan, J.S., Liu, Z.C.: Unsupervised Random Forest Indexing for Fast Action Search. In: Proc. IEEE Conf. on Computer Vision and Pattern Recognition (2011)Google Scholar
- 6.Ke, Y., Sukthankar, R., Hebert, M.: Event Detection in Crowded Videos. In: Proc. IEEE Conf. on Computer Vision (2007)Google Scholar
- 7.Derpanis, K.G., Sizintsev, M., Cannons, K., Wildes, R.P.: Efficient Action Spotting Based on a Spacetime Oriented Structure representation. In: Proc. IEEE Conf. on Computer Vision and Pattern Recognition (2010)Google Scholar
- 8.Laptev, I., Prez, P.: Retrieving Actions in Movies. In: Proc. IEEE Conf. on Computer Vision (2007)Google Scholar