Readspeaker menu

Interaction Recognition

Ali Al-Raziqi and Joachim Denzler

How do individuals interact with each other?

Project Description

Extracting compound interactions involving multiple objects is a challenging task in computer vision due different issues such as the mutual occlusions between objects,

the varying group size and issues raised from the tracker. Additionally, the single activities are uncommon compared with the activities that are performed by two or more

objects, e.g., gathering, fighting, running, etc. The purpose of this paper is to address the problemof interaction recognition among multiple objects based on dynamic

features in an unsupervised manner. Our main contribution is twofold. First, a combined framework using a tracking-by-detection framework fortrajectory extraction and

HDPs for latent interaction extraction is introduced. Second,  the introduction of a new dataset, the Cavy dataset. The Cavy dataset contains about six dominantinteractions

performed several times by two or three cavies at different locations. The cavies are interacting in complicated and unexpected ways, which leads to perform many interactions

in a short time. This makes working on this dataset more challenging.

Our framework can be seen in the figure below:


Dataset description

The Cavy dataset contains a variety of conditions that have been taken from a stationary camera. As can be observed in the below figure , sequences are recorded from different views with changing illumination and in different periods. It contains
16 sequences with 640 × 480 resolutions recorded at 7.5 frames per second (fps) with approximately 31621506 frames in total (272 GB). The sequences are recorded non-synchronously and stored in ppm

The Cavy dataset can be useful in many disciplines, in addition to computer vision, since the dataset is taken at various time, it may help the biologists to study and monitor the cavies behavior in specific periods.

Sample of frames have taken from different angles and at different time:





Ali Al-Raziqi and Joachim Denzler. Unsupervised Framework for Interactions Modeling between Multiple Objects. Proceedings of the 11th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. 509-516. 2016.

To Get access the dataset and annotated interactions Please contact :
If you use the dataset, please cite the publication mentioned above.