Piece-wise Constant Image Segmentation with a Deep Image Prior Approach
Contributo in Atti di convegno
Data di Pubblicazione:
2023
Citazione:
Piece-wise Constant Image Segmentation with a Deep Image Prior Approach / A. Benfenati, A. Catozzi, G. Franchini, F. Porta (LECTURE NOTES IN COMPUTER SCIENCE). - In: Scale Space and Variational Methods in Computer Vision / [a cura di] L. Calatroni, M. Donatelli, S. Morigi, M. Prato, M. Santacesaria. - [s.l] : Springer, 2023. - ISBN 978-3-031-31974-7. - pp. 352-362 (( Intervento presentato al 9. convegno Scale Space and Variational Methods in Computer Vision tenutosi a Santa Margherita di Pula nel 2023 [10.1007/978-3-031-31975-4_27].
Abstract:
Image segmentation is a key topic in image processing and computer vision and several approaches have been proposed in the literature to address it. The formulation of the image segmentation problem as the minimization of the Mumford-Shah energy has been one of the most commonly used techniques in the last past decades. More recently, deep learning methods have yielded a new generation of image segmentation models with remarkable performance. In this paper we propose an unsupervised deep learning approach for piece-wise image segmentation based on the so called Deep Image Prior by parameterizing the Mumford-Shah functional in terms of the weights of a convolutional neural network. Several numerical experiments on both biomedical and natural images highlight the goodness of the suggested approach. The implicit regularization provided by the Deep Image Prior model allows to also consider noisy input images and to investigate the robustness of the proposed technique with respect to the level of noise.
Tipologia IRIS:
03 - Contributo in volume
Keywords:
Image segmentation; Image segmentation with noise; Deep image prior; Neural Network; Unsupervised deep learning
Elenco autori:
A. Benfenati, A. Catozzi, G. Franchini, F. Porta
Link alla scheda completa:
Titolo del libro:
Scale Space and Variational Methods in Computer Vision