Discriminative Deep Neural Network for Predicting Knee OsteoArthritis in Early Stage - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Proceedings/Recueil Des Communications Année : 2022

Discriminative Deep Neural Network for Predicting Knee OsteoArthritis in Early Stage

Résumé

Knee osteoarthritis (OA) is a degenerative joint disease that causes physical disability worldwide and has a significant impact on public health. The diagnosis of OA is often made from X-ray images, however, this diagnosis suffers from subjectivity as it is achieved visually by evaluating symptoms according to the radiologist experience/expertise. In this article, we introduce a new deep convolutional neural network based on the standard DenseNet model to automatically score early knee OA from X-ray images. Our method consists of two main ideas: improving network texture analysis to better identify early signs of OA, and combining prediction loss with a novel discriminative loss to address the problem of the high similarity shown between knee joint radiographs of OA and non-OA subjects. Comprehensive experimental results over two large public databases demonstrate the potential of the proposed network.
Fichier non déposé

Dates et versions

hal-04559085 , version 1 (25-04-2024)

Licence

Paternité

Identifiants

Citer

Yassine Nasser, Mohammed El Hassouni, Rachid Jennane. Discriminative Deep Neural Network for Predicting Knee OsteoArthritis in Early Stage. International Workshop on Predictive Intelligence In Medicine, 13564, Springer Nature Switzerland, pp.126-136, 2022, Lecture Notes in Computer Science, ⟨10.1007/978-3-031-16919-9_12⟩. ⟨hal-04559085⟩
0 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More