Vision cognitive : apprentissage supervisé pour la segmentation d'images et de videos

Vincent Martin 1
Abstract : In this thesis we address the problem of image and video segmentation with a cognitive vision approach. More precisely, we study two major issues of the segmentation task in vision systems: the selection of an algorithm and the tuning of its free parameters according to the image contents and the application needs. We propose a learning-based methodology to easily set up and continuously adapt the segmentation task. Our first contribution is a generic optimization procedure to automatically extract optimal algorithm parameters. The evaluation of the segmentation quality is done with regards to reference segmentations. In this way, the user task is reduced to provide reference data of training images, as manual segmentations. A second contribution is a twofold strategy for the algorithm selection issue. This strategy relies on a training image set representative of the problem. The first part uses the results of the optimization stage to perform a global ranking of algorithm performance values. The second part consists in identifying different situations from the training image set and then to associate a tuned segmentation algorithm with each situation. A third contribution is a semantic approach to image segmentation. In this approach, we combine the result from the previously (bottom-up) optimized segmentations to a region labelling process. Regions labels are given by region classifiers which are trained from annotated samples. A fourth contribution is the implementation of the approach and the development of a graphical tool currently able to carry out the learning of segmentation knowledge (automatic parameter optimization, region annotations, region classifier training, and algorithm selection) and to use this knowledge to perform adaptive segmentation. We have tested our approach on two real-world applications: a biological application (detection and counting of pests on rose leaves) for the static segmentation part, and video surveillance applications for the video figure-ground segmentation part. Results, quantitative evaluations, and comparisons with non-adaptive segmentations are presented to show the potential of our approach. For the segmentation task in the biological application, the proposed adaptive segmentation approach over performs a non-adaptive segmentation in terms of segmentation quality and thus allows the vision system to count the pests with a better precision. For the figure-ground video segmentation task, the main contribution of my approach takes place at the context modelling level. By achieving dynamic background model selection based on context analysis, my approach allows to enlarge the scope of surveillance applications to high variable environments. The main limitation of our approach is its lack of adaptation to unforeseen situations. An improvement could be to use continuous learning technique to adapt the segmentation to new situations.
Document type :
Complete list of metadatas
Contributor : Vincent Martin <>
Submitted on : Tuesday, July 6, 2010 - 12:28:35 AM
Last modification on : Saturday, January 27, 2018 - 1:31:33 AM
Long-term archiving on : Thursday, October 7, 2010 - 12:18:04 PM



  • HAL Id : tel-00497797, version 1



Vincent Martin. Vision cognitive : apprentissage supervisé pour la segmentation d'images et de videos. Interface homme-machine [cs.HC]. Université Nice Sophia Antipolis, 2007. Français. ⟨tel-00497797⟩



Record views


Files downloads