Interactive Coding for Navigation in 3D scenes (GDR ISIS Grant)
This topic will be done in collaboration with Thomas Maugey and Aline Roumy (INRIA Rennes Bretagne Atlantique). The idea is to explore a new way for interactive coding of 3D scenes, in a context of free viewpoint visualization. Our collaborator have developed promising techniques (inspired by channel coding techniques) for decoding interactively multi-view videos, in function of changing points of view of the users. Our objective is to extend such techniques to geometrical data, in the context of visualization of 3D scenes, transmitted via networks.
PhaseQuantHD (ANR Grant)
High content imaging is a rapidly evolving technique which makes quantitative analysis of multiple events in a population of cultured cells. The PhaseQuantHD project aims at developing a high-content imaging system using quadriwave lateral shearing interferometry as a quantitative phase imaging modality. Automated analysis methods will be developed and optimized for this modality.
The proposed multidisciplinary consortium combines the Phasics company and three academic partners: mathematicians and biologists. MediaCoding will develop innovative automated analysis tools for supervised classification in order to optimize the information obtained from the samples.
This project will promote the development of innovative tools for biologists.
Open platform for smart and secure data management (FUI Grant)
PadDoc goal is to contribute to the acceleration of the digital transition of citizen, local and regional authorities, administrations and enterprises, by : (1) developing an open standard and innovative software and hardware resources to facilitate nearby or distant administrative formalities and procedures; (2) improving the security of the holder's personal data by putting these data under the exclusive control of the holder; (3) by exploiting unmarked communicating supports (such as smartphones or tablets) for all chain actors.
Statistical analysis of high content cellular and tissular images (SATT project)
CELL ID is a unique software solution for high content cellular imaging. It enables accurate statistical analysis on large scale cells or biological tissues databases. Off-the-shelf HCS platforms and automated microscopes daily capture and store thousands of pictures. However, downstream, cell classification is mostly done by human eye and brain (expert biologists) – a time-consuming and poorly reproducible task. CELL ID overcomes such limitations, offering the practitioner an automated supervised classification.
A low complexity system for remeshing, manipulating and visualizing high resolution 3D surface meshes (SATT Project)
The goal of GigaPolygon3D was to develop a solution allowing the visualization of out-of-core meshes constituted by billions of polygons. Contrarily to the state of the art solutions which require powerful rendering computers to deal with a huge amount of data, the proposed solution takes into account the computational capacity of the computer thanks to a GPU on-the-fly 3D mesh processing.
Compression et transmission adaptatives de maillages 3D (Young Researcher / GDR ISIS Grant)
The objective was to develop new techniques for compressing and transmitting surface meshes. More precisely, we aimed to develop an adaptive wavelet-based geometry coder, that exploits ROI (regions of interest) obtained by segmentation. This segmentation enables us to adapt the multiresolution analysis and the quantization to each segment, in function of the spatio-frequential features of the mesh geometry.
This project was done in collaboration with three members of two other CNRS labs: LSIIT (Strasbourg), and LE2I (Dijon).
French website of the project: CTAM-3D.
3D data compression technology for visualization (SATT project)
QUANTIZER is a wavelets-based compression/decompression library implementing vector quantization. Currently applied to multi-resolution 3D models, it allows to visualize multi-million triangles models on traditional Workstations, and/or multi-thousands triangles models on Smartphones. GPGPU implementation allows to use lower hardware PC resources. Culling and multiresolution technology minimize drastically GPU resources needed. It also solves the data transfer bottleneck on buses and network.
Download marketing file (pdf).
Remote visualization of volume data images (SATT project)
VOLSTREAM is a library for displaying remotely huge volume data images on PCs or mobile handsets in a client-server architecture. Compressed data lossy or lossless are streamed on demand over the networks. With the last state of the art Codec based on vector quantization, its performance allows to handle multiterabyte images with an excellent interactivity. The user is able to navigate inside the volume, increase the resolution, or change the quality up to a display without loss. Movements, changes in resolution or quality are streamed continuously. The data is not partitioned avoiding visual artifacts or abrupt consumption of resources. An implementation of the decompression on graphic card, closer to the display, save bandwidth in networks, buses, memories and all the resources upstream of visualization. Data transfers are managed by the user interaction, allowing the server to stream the just amount of compressed data needed for the visualization. This makes possible an “out-of-core” process, able to display volume data images much more heavier than the client resources. The networked mobiles and tablets can take full advantage of this innovation, which greatly reduces the data transmitted from servers to the last stage of the display.
Download marketing file (pdf).
Analyse Multirésolution d’Objets 3D Animés (Young Researcher / GDR ISIS Grant)
The goal of this project was the design of multiresolution analysis for 3D animations. The starting idea was to extend wavelet-based techniques for video and surfaces to sequences of meshes with temporal coherency, widely used to represent animated 3D objets. The main application was geometry compression.
This project was done in collaboration with three members of three other CNRS labs: LSIIT (Strasbourg), GIPSA (Grenoble), LJK (Grenoble).
French website of the project AMOA-3D.