<!-- *** Custom HTML *** --> Purpose: The automatic segmentation of multiple sclerosis lesions in magnetic resonance imaging has the potential to reduce radiologists’ efforts on a daily time-consuming task and to bring more reproducibility. Almost all new segmentation techniques make use of convolutional neural networks with their own different architecture. Architectural choices are rarely explained. We aimed at presenting the relevance of a U-net-like architecture for our specific task and at building an efficient and simple model. Approach: An experimental study was performed by observing the impact of applying different mutations and deletions to a simple U-net-like architecture. Results: The power of the U-net architecture is explained by the joint benefits of using an encoder–decoder architecture and by linking them with long skip connections. Augmenting the number of convolutional layers and decreasing the number of feature maps allowed us to build an exceptionally light and competitive architecture, the minimally parameterized U-net (MPU-net), with only ∼30,000 parameters. Conclusion: The empirical study of the U-net has led to a better understanding of its architecture. It has guided the building of the MPU-net, a model far less parameterized than others (at least by a factor of seven). This neural network achieves a human-level segmentation of multiple sclerosis lesions on fluid-attenuated inversion recovery images only. It shows that this segmentation task does not necessitate overly complicated models to be achieved. This gives the opportunity to build more explainable models that can help such methods to be adopted in a clinical environment.

With the emergence of huge volumes of high-resolution Hyperspectral Images (HSI) produced by different types of imaging sensors, analyzing and retrieving these images require effective image description and quantification techniques. Compared to remote sensing RGB images, HSI data contain hundreds of spectral bands (varying from the visible to the infrared ranges) allowing profile materials and organisms that only hyperspectral sensors can provide. In this article, we study the importance of spectral sensitivity functions in constructing discriminative representation of hyperspectral images. The main goal of such representation is to improve image content recognition by focusing the processing on only the most relevant spectral channels. The underlying hypothesis is that for a given category, the content of each image is better extracted through a specific set of spectral sensitivity functions. Those spectral sensitivity functions are evaluated in a Content-Based Image Retrieval (CBIR) framework. In this work, we propose a new HSI dataset for the remote sensing community, specifically designed for Hyperspectral remote sensing retrieval and classification. Exhaustive experiments have been conducted on this dataset and on a literature dataset. Obtained retrieval results prove that the physical measurements and optical properties of the scene contained in the HSI contribute in an accurate image content description than the information provided by the RGB image presentation.

Recent developments in hyperspectral sensors allow us to obtain high spectral and spatial resolutions that are close to the optical and physical structures of acquired surfaces. Consequently, hyperspectral imaging is used for its potential gain of accuracy. To preserve this metrological potential, generated bias, errors, and uncertainties must be managed at all subsequent processing levels. Based on the argument that a spectral image processing should avoid the linear approach, this paper proposes several protocols for assessing the quality of spectral ordering relations. The protocols include considerations of theoretical properties to ensure result stability, in physical aspects of spectral processing to ensure the link to physical properties of materials, and experimental results using spectral images with physical ground truth to assess bias and errors. Full-band ordering relations are compared in order to find that which satisfies all of the expected properties of a metrological spectral image processing.

In order to model the concept of emotion and to extract the emotional impact from images, one may search suitable image processing features. However, in the literature, there is no consensus on the ones to consider since they are often linked to the application. Obviously, the perception of emotion is not only influenced by the content of the images, it is also modified by some personal experiences like cultural aspects and semantic associated to some colours or objects. In this paper, we choose low level features frequently used in CBIR especially those based on SIFT descriptors. To take into account the complex process of emotion perception, we also consider colour and texture features and one global scene descriptor: GIST. We supposed the chosen features could implicitly encode high-level information about emotions due to their accuracy in the different CBIR applications of the literature. We test our methodology on two databases: SENSE and IAPS.

Although a lot of work has been done on optical coherence tomography and color images in order to detect and quantify diseases such as diabetic retinopathy, exudates, or neovascularizations, none of them is able to evaluate the diffusion of the neovascularizations in retinas. Our work has been to develop a tool that is able to quantify a neovascularization and the fluorescein leakage during an angiography. The proposed method has been developed following a clinical trial protocol; images are taken by a Spectralis (Heidelberg Engineering). Detections are done using a supervised classification using specific features. Images and their detected neovascularizations are then spatially matched by an image registration. We compute the expansion speed of the liquid that we call diffusion index. This last one specifies the state of the disease, permits indication of the activity of neovascularizations, and allows a follow-up of patients. The method proposed in this paper has been built to be robust, even with laser impacts, to compute a diffusion index.

Glial tumors represent the leading etiology of primary brain tumors. Their particularities lie in (i) their location in a highly functional organ that is difficult to access surgically, including for biopsy, and (ii) their rapid, anisotropic mode of extension, notably via the fiber bundles of the white matter, which further limits the possibilities of resection. The use of mathematical tools enables the development of numerical models representative of the oncotype, genotype, evolution, and therapeutic response of lesions. The significant development of digital technologies linked to high-resolution NMR exploration, coupled with the possibilities offered by AI, means that we can envisage the creation of digital twins of tumors and their host organs, thus reducing the use of physical sampling.