Geometric-based segmentation of polarization-encoded images

International audience

Physical Processing of Polarization Images Noise Estimation of Polarization-Encoded Images by Peano-Hilbert Fractal Path

An SVD-Based Approach for Ghost Detection and Removal in High Dynamic Range Images

International audience; In this paper, we propose a simple method for the ghost detection problem in the context of merging multiple low dynamic range (LDR) images to form a high dynamic range (HDR) image. We show that the second biggest singular values extracted over local spatio-temporal neighbourhoods can be effectively used for ghost region detection. Furthermore, we combine the proposed method with an exposure fusion technique to generate final HDR image free of ghosting artefacts. We present experimental results to illustrate the efficiency of the proposed method and quantitative comparison with other existing approaches show the good performance of our method in detecting and removin…

Can SegFormer be a True Competitor to U-Net for Medical Image Segmentation?

The U-Net model, introduced in 2015, is established as the state-of-the-art architecture for medical image segmentation, along with its variants UNet++, nnU-Net, V-Net, etc. Vision transformers made a breakthrough in the computer vision world in 2021. Since then, many transformer based architectures or hybrid architectures (combining convolutional blocks and transformer blocks) have been proposed for image segmentation, that are challenging the predominance of U-Net. In this paper, we ask the question whether transformers could overtake U-Net for medical image segmentation. We compare SegFormer, one of the most popular transformer architectures for segmentation, to U-Net using three publicl…

Spectral Clustering of Shape and Probability Prior Models for Automatic Prostate Segmentation in Ultrasound Images

International audience; Imaging artifacts in Transrectal Ultrasound (TRUS) images and inter-patient variations in prostate shape and size challenge computer-aided automatic or semi-automatic segmentation of the prostate. In this paper, we propose to use multiple mean parametric models derived from principal component analysis (PCA) of shape and posterior probability information to segment the prostate. In contrast to traditional statistical models of shape and intensity priors, we use posterior probability of the prostate region determined from random forest classification to build, initialize and propagate our model. Multiple mean models derived from spectral clustering of combined shape a…

Filtering of polarization-encoded images by Piano-Hilbert fractal path

International audience

Scanning From Heating: 3D Shape Estimation of Transparent Objects from Local Surface Heating

International audience

An optimization approach to segment breast lesions in ultra-sound images using clinically validated visual cues

International audience; As long as breast cancer remains the leading cause of cancer deaths among female population world wide, developing tools to assist radiologists during the diagnosis process is necessary. However, most of the technologies developed in the imaging laboratories are rarely integrated in this assessing process, as they are based on information cues differing from those used by clinicians. In order to grant Computer Aided Diagnosis (CAD) systems with these information cues when performing non-aided diagnosis, better segmentation strategies are needed to automatically produce accurate delineations of the breast structures. This paper proposes a highly modular and flexible f…

Toward an International Curricula Network for exchanges and LifeLong Learning

4 pages; International audience; During the last years, an important work has been achieved within the framework of the European programs in order to emphasize and allow the mobility of students, the exchange of teachers, the recognition of diploma (programs ERASMUS, now in the LLP program)... These reflections have been done for classical programs within HEIs throughout Europe and also within the programs dedicated to vocational studies (LEONARDO, ValeurTech, EQF) to emphasize Life Long Learning and credit accumulation all around the life, with an approach based on competences [1]. The main purposes in these programs can be summarized as follows: − Favoring the equality among citizens in E…