Menu

Francisco LAMOSA

COLCHESTER

En résumé

Spécialiste en systèmes d’imagerie et traitement/analyse d’images orienté vers le client avec base solide en développement pour les processeurs CPU et GPU en C et C++, familier avec les caméras et cartes d’acquisition. Expérimenté dans l’identification et documentation des besoins des clients, le présentation des solutions, le développement et introduction des systèmes (incluant l’entraînement et le support des clients). Expérience internationale dans des organisations allant de PME aux multinationales.

I like working with team members, clients, suppliers to understand requirements and provide support/ training to ensure clients make informed choices. In this process, I have delivered image processing solutions,using both CPUs and GPUs, to clients' satisfaction, managing the requirements of the different stakeholders, starting with the system requirements to design and implementation to integration , and finally to providing training and technical support.

Having lived on 3 continents and being multilingual, I can in operate in different environments

Mes compétences :
Image processing
Systems integration
Cuda
Opencv
Mil
Matlab
C++
OpenCL
C class
Formation professionnelle
Computer vision

Entreprises

  • De Beers Technologies - Senior Image Processing Engineer

    2018 - maintenant -Colour grade classification using machine learning techniques
    -Point cloud segmentation of OCT backscattered signals

  • Continental - Senior Algorithm Engineer

    Toulouse 2016 - 2018 Implementation of SfM algorithms on Broadcom Pixel processor (embedded video processor) with ARM host
    KLT optical flow components were implemented on hardware: feature(corner) detect, matching/tracking for object detection for vehicle collision avoidance (as part of an ADAS system).

    Implementation des elements d'algorithme SfM sur processeur embarque de Broadcom (processeur video embarque) avec microcontrolleur ARM
    Element du flow optique sur le processeur: detecteur Harris, HoG - tracking

  • OMRON - Senior Research Engineer

    ROSNY SOUS BOIS 2015 - maintenant Investigated noise reduction algorithms (with Matlab, C++, CUDA)
    Implementation of noise filtering in CUDA
    Assumed responsibility for maintenance of visualization module
    Segmentation of point clouds for registration with CAD models
    Kyoto, Japan

  • Dage Precision Industries - Senior Software Engineer

    2013 - maintenant -Designed modules for implementing GPU processing components of X-ray inspection system used in semiconductor inspections, integration, and validation
    -Investigating algorithms for image processing analysis components (feature detection, matching, and image stitching)
    -Meeting with external contributors to coordinate development efforts
    -Work with clients in Japan to understand needs and provide viable solution

  • AMD - Software engineer

    Californie 2011 - 2012 Developed tools for tracing OpenCL API calls– one significant feature was GPU occupancy – found in AMD APP Profiler SDK.
    Contributed to testing framework: wrote unit tests using Google testing framework; introduced use of Google Mock framework to test new features when software stack not yet available
    Designed, implemented, and used an OpenCL-based framework for validating hardware debugging features of SI family GPUs.
    Wrote core trap handler (using GPU ISA) for the h/w debugger to process breakpoint exceptions
    Contributed to resource planning and risk estimation of hardware debugger project

  • SPG 3D - Developed coded projection algorithm for feature matching for 3-d reconstruction.

    2010 - 2011 Evaluated tools to assist in development process: UML, prototyping environments (Scilab, Python, Octave), image processing libraries and packages; developed guidelines standards for software
    Tested approaches to improve structured light stereo under noisy image conditions (low intensity light, rough surface with highly variable reflectance, limited depth of focus)
    Calibrated camera-projector system
    Delivered C++ coded projection-based feature matching library for 3-D reconstruction; basic functionality provided by CImg class template library with in-house extension (C++ based)

  • Vision4ce Limited. - Developed CUDA algorithms for image processing.

    2009 - 2010 Developed image processing algorithms on NVIDIA GPU using CUDA runtime and driver
    CUDA API for unsharp masking, contrast enhancement, and enhancing image stabilization
    algorithm; prototyping done in Matlab
    Established performance measures with respect to CPU-based, optimized algorithms
    Integrated these algorithms into C++ based software libraries (with a C API)
    Wrote application notes and documentation for the new functionality

  • Computer Recognition Systems, Ltd. - Software development (C and C++)

    2002 - 2004 Developed functionality for data transfer between systems, including encryption, architecture for a traffic analysis system.

  • Matrox Electronic Systems - Applications specialist/software development specialist

    1999 - 2002 Provided telephone and on-site customer support. Investigated customer needs and provided sample code to address the client problem. When necessary, escalated bugs to appropritate software/hardware teams. Developed configuration files to interface cameras with frame grabbers. Developed algorithms to address client problems.

  • Noesis Vision Inc. - Applications engineer - image processing.

    1997 - 1999 Developed turnkey applications for customers, including algorithms, software system (including user interface), integration with control and lighting systems, delivery and installation, training, and post-delivery support. I also worked on presales suport, by discussing needs and providing sample image processing scripts to demonstrate potential solutions.

Formations

  • Kumamoto University (Kumamoto)

    Kumamoto 2004 - 2009 PhD.

    Computer Science - Doctorate in computer vision: thesis was in the fusion of multi-camera data using perspective re-projection of 3 different camera scenes onto a common frame (using inverse perspective mapping). Applied technique to road traffic monitoring to reduce the impact of occlusions and lighting variability.

Réseau

Annuaire des membres :