HAL Portal UTC Université de Technologie de Compiègne
Not a member yet
11652 research outputs found
Sort by
Application of belief functions to medical image segmentation: A review
No full textInternational audienceThe investigation of uncertainty is of major importance in risk-critical applications, such as medical image segmentation. Belief function theory, a formal framework for uncertainty analysis and multiple evidence fusion, has made significant contributions to medical image segmentation, especially since the development of deep learning. In this paper, we provide an introduction to the topic of medical image segmentation methods using belief function theory. We classify the methods according to the fusion step and explain how information with uncertainty or imprecision is modeled and fused with belief function theory. In addition, we discuss the challenges and limitations of present belief function-based medical image segmentation and propose orientations for future research. Future research could investigate both belief function theory and deep learning to achieve more promising and reliable segmentation results
Non woven electrospun scaffold with continuous gradient from honeycomb like to aligned structures for osteotendinous junction tissue engineering
No full textInternational audienc
Process-dependent material characteristics of DMLS-manufactured specimens
No full textInternational audienc
Ridge extraction to identify activated motor unit on heatmap generated by source localization of HD-sEMG signals
No full textInternational audienceThis research explores the utilization of ridge extraction techniques for the identification of activated motor units from heat map images generated by source localization of high density surface electromyogram (HD-sEMG) signals. The primary objective of this research is to accurately determine the number of activated motor units in these images, facilitating enhanced understanding of neuromuscular activities. The study involves the application of two ridge detection approaches: the first employs Otsu's thresholding in conjunction with the watershed algorithm, while the second utilizes the Hessian matrix. By applying these methods, the aim is to enhance the accuracy of identifying ridges associated with activated motor units in the heat map images. Following ridge extraction, two methods are employed to estimate the number of motor units within the images. The first method employs the K-Nearest Neighbors (KNN) algorithm, demonstrating promising accuracy in determining the count of motor units. The second method involves the utilization of K-Means clustering; however, its performance is observed to be less consistent, especially when confronted with images containing more than two motor units. These findings underscore the importance of selecting appropriate analysis methods when working with complex neuromuscular data
High Precision Machining of a Displacement Sensor for Helicoidal Motions
No full textInternational audienceThis research study presents the design and the high precision manufacture procedure of a fiber-optic displacement sensor. It is composed of two fiber-optic probes associated with a structure of a cones’ grating. The sensor is characterized by its ability to measure the linear displacement for an axis performing a helicoidal motion. This motion has been demonstrated on a high precision lathe; where the spindle provided the rotational motion, associated to a translational motion on the linear stage. This allowed to obtain the two simultaneous motions. The displacement of the translational stage is measured by the sensor in real time. Firstly, a highly precise geometric model of the reflector part for the sensor was developed. This model provided a specific geometry for the cones-assembled grating, which has been precisely manufactured. The geometric parameters and the surface characteristics of each step in the fabricated grating were both identified in situ on the lathe. The agreement between simulation and experimental results is excellent. The performances of the fiber-optic displacement sensor were identified in-situ on the lathe. The analysis of the voltage output signals from the two fiber-optic probes is used to measure the grating displacement. The unbalanced rotation due to non-centered axes was also characterized. The sensor provided a micrometric resolution, on a measurement range of more than one centimeter
Deep Learning Approach for Hand Gesture Recognition: Applications in Deaf Communication and Healthcare
No full textInternational audienc
Using a Reduced-Order Modeling strategy for accounting for texture information in the context of polycrystalline plasticity
No full textInternational audienc
Novel and legacy brominated flame retardants in snakes and frogs: Tissue distribution, biomagnification, and maternal transfer
No full textInternational audienc
Artistes vs industries culturelles, culture libre
No full textIntervention dans le cadre du colloque « Cliquer l’archive - le numérique et la sauvegarde des pratiques artistiques », tenu à l'École nationale supérieure des beaux-arts de Paris, les 12 et 13 décembre 2023