وزارة التعليم العالي و البحث العلمي
Research centre in Industrial Technologies -CRTI-
EChahid Mohammed ABASSI
Signal processing and imagery division
Images segmentation and analysis
- Naim RAMOU Senior Researcher A Team leader
- Yamina BOUTICHE Senior Researcher B
- Nabil CHETIH Senior Researcher B
- Mohammed KHORCHEF Research fellow
Hybrid segmentation approach for the analysis of digital images used in different applications. [2022-2024]
his project focuses on hybrid image segmentation in a Bayesian framework for different applications. The first proposed model combines the Dirichlet process (DP) and the Mumford shah based model level set. For an image, it is divided into homogeneous regions and similar regions between images are grouped into classes. On the one hand, thanks to the DP, it is not necessary to define a priori the number of regions per image and the number of classes, common or not. On the other hand, the level set ensures spatial homogeneity. Although it favours the creation of classes, the Dirichlet process alone is not suitable for image segmentation. Indeed, the spatial interactions inherent to an image are not taken into account. It is therefore necessary to combine it with another model, for example level set, for spatial homogeneity in the segmented image. To make a comparative study between segmentation methods, we are interested in other segmentation techniques, which are useful in different applications. For example, in agronomy or in forest fire detection. To address these issues in this project we will focus on three types of image processing and segmentation:1. Hybridization between variational and Bayesian methods.2. Algorithms that combine image processing based on colour indices and shape with a comparison to deep learning.3. Supervised and unsupervised segmentation techniques. Application 1:The detection of forest fires plays an important role in preventing the damage caused by their spread. The analysis of video information is an interesting and very promising option for real-time monitoring and detection of forest fires. Image processing algorithms are the heart of a detection system. Segmentation is the crucial and determining step in the processing chain. The objective of this application is the design of a fire monitoring and detection system for the minimisation of damage by timely intervention of civil protection elements. Application 2: The basic idea of this application is to have information from the segmentation of agronomic images in order to help farmers make decisions that will allow them to increase their productivity. The application we plan to develop will be simple to use. This application will bring together the functions of image segmentation using the different techniques. This will provide useful information to help farmers make decisions to increase and monitor their productivity.
Main menu :
Newsletter :
Social networks :
