Software research and development in the 3D scanner field
April 2005 - Current, Japanese company manufacturing 3D scanners

This is my second experience in the 3D scanner field. Before my arrival, the company was focused on hardware. I have been in charge to develop new software, especially in the 3D model measurement field. As it is a small company, I can handle a range of responsibilities and I appreciate in particular the freedom I have with regard to the way to reach the objectives.

I developed several software:
Measurements on 3D models
Automatic measurement list editor
Scan anonymization
Scan stitching (semi-automatic scan fusion)
Company scanner data visualizer
And also some little programs for company internal use

Those developments required to evaluate and use existing methods, but also to make research about new approaches. That permitted me to increase my technical skills particularly in the following fields:
Image and depth map (DEM) operations
Filtering
Mathematical morphology
Vectorization, operations on vectorized curves
Model-based object detection
Lens distortion measurement, calibration and correction
Data structures
Quadtree
Octree
kd-tree and R-tree
Various structures for acceleration of mesh operations
Mesh operations
Reconstruction
2-manifold correction
Hole filling
Merging
Remeshing
Simplification, Level-of-detail
Comparison, positioning
Texture projection
Software copy protection system
Reading/writing of various 2D/3D and XML file formats
TCP and UDP communication between computers

The contacts with clients helped me to improve my knowledge of the market. I have also been responsible to provide software documentation, demonstration and training.

Software were developed using C++ language and the OpenGL library.

Development of software for visualization and processing of volumetric images
March - September 2004, Pierre Fabre Dermo-Cosmétique - http://www.pierre-fabre.com

Within the framework of various research projects, software of visualization and processing of volumetric images had become necessary. Although initially intended to be used with magnetic resonance images, this software has been adapted to data coming from other imaging techniques: confocal microscopy, echography, optical coherence tomography.

By performing this project, I have become familiar with the following techniques:
Reading of DICOM files
Methods of volumetric display:
Interactive rendering: the software must display interactive representation of the volumetric data, so several methods were applied (succession of 2D textures, Shear-Warp Factorization, cloud of dots)
High quality rendering: Ray Casting
Volumetric image processing
Extraction of vascularization network and their polygonation
Active contour models for accurate location of vascularization network
Design of an evolutionary user interface

During my work, the Centre Jean-Louis Alibert obtained an ISO 9001 certification. This has spurred on a greater standardization of the process of software development in which I took part.

The software was developed in C++ language, using the OpenGL library and also integrating Matlab functions for the processing of some data.

Screenshots  


Development of a software of control for a 3D imager
January - September 2003, Pierre Fabre Dermo-Cosmétique - http://www.pierre-fabre.com

During 8 months, I was in charge of the development of new control software for a prototype of three-dimensional imager. Used in the dermatological field, it digitizes body shape and provides an exploitable 3D model for further analysis.

With this very enriching work, I used and improve my skills in the following fields:
Calibration of a vision system
Processing of profile images coming from four profilometers (with a method giving a sub-pixel precision)
Construction of a 3D mesh from a group of dots (in particular by a method derived from the Marching-Cubes)
Colorimetric correction of the texture images
Mapping of the texture images on the 3D mesh
Design of a user interface

During this work, I performed researches on experimental meshing method partly based on volumetric mathematical morphology. An article was published in the framework of the RFIA 2004 (Shape Recognition and Artificial Intelligence Francophone Congress) in collaboration with the LAAS-CNRS (Laboratory for Analysis and Architecture of Systems - National Center for Scientific Research).

The development of the application was made using C++ language, always keeping in mind later reusability of the data structures management and treatment libraries.

Screenshots  


Test-bench for surface potential measurement
April - June 1999, Laboratoire de Génie Electrique de Toulouse - http://www.lget.ups-tlse.fr

In order to characterize certain electrical insulating materials, a bench for measuring the surface potential was implemented by the Electric Engineering Laboratory of Toulouse. My job was to finalize the prototype, which led me to develop electronic cards allowing computer to control operations. I also had to develop driving software supervising the measurements results:
Mapping of the charge on the surface of a material
Analysis of temporal decrease of these charges at a given point

This software was developed using Microsoft Visual Basic.

           Description of my projets (PDF file)        Description of my projets (Word file)