Conference Paper

Real-Time Segmentation of Moving Objects in H.264 Compressed Domain with Dynamic Design of Fuzzy Sets.

Escuela Universitaria de Ingenieria Tecnica Industrial, Universidad de Castilla-La Mancha Toledo, Spain
Conference: Proceedings of the Joint 2009 International Fuzzy Systems Association World Congress and 2009 European Society of Fuzzy Logic and Technology Conference, Lisbon, Portugal, July 20-24, 2009
Source: DBLP


This paper presents a real-time segmentation algo-rithm to obtain moving objects from the H.264 compressed domain. The proposed segmentation works with very little information and is based on two features of the H.264 compressed video: motion vec-tors associated to the macroblocks and decision modes. The algo-rithm uses fuzzy logic and allows to describe position, velocity and size of the detected regions in a comprehensive way, so the proposed approach works with low level information but manages highly com-prehensive linguistic concepts. The performance of the algorithm is improved using dynamic design of fuzzy sets that avoids merge and split problems. Experimental results for several traffic scenes demonstrate the real-time performance and the encouraging results in diverse situations. Keywords— Moving object detection, image segmentation, H.264 advanced video coding, dynamic fuzzy sets.

Download full-text


Available from: Juan Moreno García, Jun 26, 2014
  • [Show abstract] [Hide abstract]
    ABSTRACT: A new algorithm to classify moving objects in monitored environments is presented. The approach is based on a supervised machine learning algorithm and uses as input data the results obtained in a previously developed segmentation algorithm. The algorithm has a training stage which uses clustering algorithms and a learning stage to learn the features of each kind of object and to be able to classify moving objects in video scenes. The labelling approach is focused on video-surveillance monitoring, thus it runs in real-time, and it has been designed to exploit different features of the objects: position, size, shape and motion. Experimental results show promising performance in terms of both accuracy and efficiency.
    No preview · Conference Paper · Jan 2010