Image Processing Group
The main topics investigated at the Image Processing group are character recognition, color segmentation, text localization, default detection with supervised classification, medical image analysis, modelisation and segmentation.
In addition to the basic techniques, we use and develop methods based on feature extraction and selection, neural networks, support vector machines, wavelet transforms, Gabor filters, snake and active contours, fast marching, and watershed.
We work, or have recently worked, on fruits classification, on portable automatic reading system, on emboly detection from medical images, and tumor segmentation.
Applications : on-line sorting of fruits by industrial vision (RW CAPA project), automatic reading system for visually impaired and blind people (RW SYPOLE project), computer-aided diagnosis of pulmonary embolism (RW iMed project), radiotherapy planning and monitoring tools (RW Mercator project).