Skip to content

Implementation of a Cross Dataset framework that both detects actions in a crowded scenes and adapt Models from a Source datset to a new Target dataset with minumum annotaions needed.

Notifications You must be signed in to change notification settings

aelnouby/Cross-Dataset-Action-Detection

Repository files navigation

Cross Dataset Action Detection

#Abstract In recent years, many research works have been carried out to recognize human actions from video clips. To learn an effective action classifier, most of the previous approaches rely on enough training labels. When being re- quired to recognize the action in a different dataset, these approaches have to re-train the model using new labels. However, labeling video sequences is a very tedious and time-consuming task, especially when detailed spatial lo- cations and time durations are required. In this paper, we propose an adaptive action detection approach which reduces the requirement of training labels and is able to handle the task of cross-dataset action detection with few or no extra training labels. Our approach combines model adaptation and action detection into a Maximum a Posterior (MAP) estimation framework, which explores the spatialtemporal coherence of actions and makes good use of the prior information which can be obtained without supervision. Our approach obtains state-of-the-art results on KTH action dataset using only 50% of the training labels in tradition approaches. Furthermore, we show that our approach is effective for the cross-dataset detection which adapts the model trained on KTH to two other challenging datasets.

##Adaptation Results :

Using alpha = 0.1 in the mean adaptaion equation and executing predict_target.py, the results were as follows

Adaptation Microsoft Training Microsoft Testing
No Adaptaion (pure KTH mode) 37.2% 46.62%
1st Iteration 77.3% 63.2%
2nd Iteration 85.1% 67.85%
3nd Iteration 94.5% 71.7%

##Detection Results :

#Datasets I. Recognition of human actions, KTH. [Data]

II. MSR Action Dataset I [Data]

#Software Subvolume Branch-and-Bound Search binaries(WIN32) [binaries]

In Linux use wine as follows :

wine SubVolumeSearch.exe infile outfile numclasses width height penaltyValue classthreshold1 classthreshold2 ... classthresholdn

##Refrences [1] Cao, Liangliang, Zicheng Liu, and Thomas S. Huang. "Cross-dataset action detection." Computer vision and pattern recognition (CVPR), 2010 IEEE conference on. IEEE, 2010. [PDF]

[2] Yuan, Junsong, Zicheng Liu, and Ying Wu. "Discriminative video pattern search for efficient action detection." Pattern Analysis and Machine Intelligence, IEEE Transactions on 33.9 (2011): 1728-1743.[PDF]

[3] Project page by Liangliang Cao [Page]

About

Implementation of a Cross Dataset framework that both detects actions in a crowded scenes and adapt Models from a Source datset to a new Target dataset with minumum annotaions needed.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages