Altran

About: Altran is a company organization based out in Neuilly-sur-Seine, France. It is known for research contribution in the topics: Software development & Formal verification. The organization has 488 authors who have published 512 publications receiving 6395 citations.

A Deep Learning Approach for Estimation of the Nearshore Bathymetry

Abstract: Benshila, R.; Thoumyre, G.; Al Najar, M.; Absessolo, G.; Almar, R.; Bergsma, E.; Hugonnard, G.; Labracherie, L.; Lavie, B.; Ragonneau, T.; Simon, E.; Vieuble, B., and Wilson D., 2020. A deep learning approach for estimation of the nearshore bathymetry. In: Malvarez, G. and Navas, F. (eds.), Global Coastal Issues of 2020. Journal of Coastal Research, Special Issue No. 95, pp. 1011-1015. Coconut Creek (Florida), ISSN 0749-0208.Bathymetry is an important factor in determining wave and current transformation in coastal and surface areas but is often poorly understood. However, its knowledge is crucial for hydro-morphodynamic forecasting and monitoring. Available for a long time only via in-situ measurement, the advent of video and satellite imagery has allowed the emergence of inversion methods from surface observations. With the advent of methods and architectures adapted to big data, a treatment via a deep learning approach seems now promising. This article provides a first overview of such possibilities with synthetic cases and its potential application on a real case.

Accuracy analysis of an augmented reality system

Abstract: Work described in this contribution focuses on error analysis in augmented reality (AR) systems. The tracking, the process of locating an object (e.g. fiducial marker) in an environment, is critical to the accuracy of AR applications as more realistic results can be obtained in the presence of accurate AR registration. This deals with 2D error estimations of the edge detection process, the starting step of the whole tracking procedure enabling to determine the outline of the imaged marker. Using fitting techniques to describe the geometric features composing the outline, errors bounds are determined and, as a result of this step, edge detection errors are estimated. These 2D errors are then propagated up to the final tracking step.

Identification and characterization of internal waves in SAR images along the coast of Norway

Abstract: The Synthetic Aperture Radars (SARs) on ERS and RADARSAT are sensitive to changes in the small-scale surface roughness present on the ocean surface, which is altered by the velocity field associated with internal waves. More than 2600 SAR images from Norwegian waters were analyzed to obtain geometric and radiometric signature statistics of internal waves. The statistical analysis shows that the dispersion properties can be described by a single wavelength scaling parameter of 1.4, allowing internal waves to be detected and characterized with a wavelet based algorithm. Internal waves are identified mainly in four “hot-spot” areas with somewhat distinctive characteristics. Information on internal waves is useful for the offshore oil industry and for the military, both for planning purposes and operationally in near real time.

Framework for augmented reality in Minimally Invasive laparoscopic surgery

Abstract: This article presents a framework for fusing preoperative data and intra-operative data for surgery guidance. This framework is employed in the context of Minimally Invasive Surgery (MIS) of the liver. From stereoscopic images a three dimensional point cloud is reconstructed in real-time. This point cloud is then used to register a patient-specific biomechanical model derived from Computed Tomography images onto the laparoscopic view. In this way internal structures such as vessels and tumors can be visualized to help the surgeon during the procedure. This is particularly relevant since abdominal organs undergo large deformations in the course of the surgery, making it difficult for surgeons to correlate the laparoscopic view with the pre-operative images. Our method has the potential to reduce the duration of the operation as the biomechanical model makes it possible to estimate the in-depth position of tumors and vessels at any time of the surgery, which is essential to the surgical decision process. Results show that our method can be successfully applied during laparoscopic procedure without interfering with the surgical work flow.