Abstract | ||
---|---|---|
In this paper, we propose a novel model for activity detection in videos via discriminative patches. Each frame is represented as a set of mid-level discriminative patches, which are extracted automatically by association rule mining with convolutional neural networks (CNN) activations. Based on the observation that there are more discriminative patches occurring in the climax stages of an activity, we propose a climax stages discovering method, where the climax of an activity is defined as the continuous frames which have more discriminative patches. Then the patches are further purified by a punishment rule, which ensures that the discriminative patches seldom occur in non-climax stages. Furthermore, the refined patches are used for temporal detection on continuous execution. We demonstrate state-of-the-art performance for activity detection on un-segmented UT-Interaction Set #1. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1145/3007669.3007720 | ICIMCS |
Keywords | Field | DocType |
Activity detection, Discriminative patches, Climax stages | Pattern recognition,Convolutional neural network,Computer science,Speech recognition,Association rule learning,Activity detection,Artificial intelligence,Discriminative model | Conference |
Citations | PageRank | References |
0 | 0.34 | 2 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Dingyi Shan | 1 | 0 | 0.68 |
Laiyun Qing | 2 | 337 | 24.66 |
Jun Miao | 3 | 220 | 22.17 |