ABSTRACT
Deep venous thrombosis (DVT) is an inappropriate formation of a thrombus in a deep vein. Three physio-pathological mechanisms can contribute, isolated or combined, to the development of a DVT: venous stasis, endothelial injury and hypercoagulability. The main aim of our project is to characterize the structure of DVT in order to identify one or more factors responsible for its formation. In this project, we developed feature extraction, identification and classification approaches based on scattering operators and statistical methods in order to characterize DVT or phlebitis. It is worth highlighting that the detachment and migration of the thrombosis formed in a vein from the lower extremities to the pulmonary arteries can cause sudden partial or total obliteration of the artery. This obliteration is identified as a major complication called Pulmonary Embolism. In our project, we are looking for the relationship between DVT epidemiology and the thrombus structure. We extract features from ultrasound images using a scattering operator and high-order statistics. Then, the obtained features are analyzed using several classification technics to find the main cause of the DVT or the presence of PE. Experimental results are presented and discussed.
