Handling Uncertainty for Video Event Recognition

Abstract : This paper presents a cognitive vision approach for video event recognition able of handling the uncertainty of the recognition process. The recognition task is complex because of image noise, of segmentation and classification issues. In this work, we extend the event recognition algorithm (crisp algorithm) proposed in [1] by proposing a geometric method which handles the uncertainty of the recognition process. This method consists in computing the precision of the 3D information of the mobile objects evolving in the scene for each frame of the video sequence. We use the computed information to calculate the probability of the event. The proposed method is tested with videos of everyday activities of elderly people. Events of interest have been modeled with the help of medical experts (i.e. gerontologists). The experimental results show that the proposed approach improves significantly the process of recognition and can characterize the likelihood of the recognized events.
Complete list of metadatas

https://hal.inria.fr/inria-00502680
Contributor : Rim Romdhane <>
Submitted on : Thursday, July 15, 2010 - 2:35:41 PM
Last modification on : Tuesday, July 24, 2018 - 3:48:06 PM
Long-term archiving on : Friday, October 22, 2010 - 12:08:47 PM

File

ICDParticleFinalVersion.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : inria-00502680, version 1

Collections

Citation

Rim Romdhane, François Bremond, Monique Thonnat. Handling Uncertainty for Video Event Recognition. ICDP 2009, Dec 2009, London, United Kingdom. ⟨inria-00502680⟩

Share

Metrics

Record views

250

Files downloads

193