Aerial images segmentation - 2017
Technologies: Python, TensorFlow, CNN
TeraDeep: Implement and train a convolutional neural network for a task of building detection through pixel segmentation (click here to see an example of result).
This website is just a simple portfolio where I expose some of my personal projects and some works I developed during my studies and internships.
Some of those projects are partially or completely available on my github page.
Technologies: Python, TensorFlow, CNN
TeraDeep: Implement and train a convolutional neural network for a task of building detection through pixel segmentation (click here to see an example of result).
Technologies: C++, Lua, Torch, Python, Flask
TeraDeep: For this project, I was in charge of developing a web server to integrate and deploy deep learning services for diverse tasks of image and video recognition. I also had to design the associated API.
Technologies: Python, TensorFlow, Seq2seq, Django
Personal Project: My attempt to reproduce A Neural Conversational Model, implementing a 2 layers LSTM encoder/decoder. All the code is available here.
Technologies: Python, TensorFlow
Personal Project: Attempt to create and experiment different models for music composition. You can listen a generated sample here.
Technologies: Python, Numpy, RNN, NLP
Personal Project: Implementation from scratch of the stanford Recursive Neural Tensor Network to detect the sentiment of movies critics.
Technologies: Python, Theano, Torch, CNN
School Project: Apply convolutional neural network on graphs for 3d meshes segmentation and classification.
Technologies: Spark, Rdd, SQL, MapReduce
School Project: Optimising SQL queries with Spark.
Technologies: Dart, Polymer, Plotly.js
Personal Project: Creation of a web application which extract my IMDb ratings to do some data visualisation and compute some statistics.
Technologies: C++
School Project: Implementation of a path tracing algorithm. Based on the Nori frameworks
Technologies: Python
School Project: Implementation of the bellmans equation to guide a mouse in a continious world.
Technologies: Matlab
School Project: Tuning a neural networks for image classification and comparison with other techniques.
Technologies: Java, Android SDK
Personal Project: Record and keep track of our performances with this simple app.
Technologies: Unity3d, C#
School Project: Creation of a simple VR app for Google Cardboard. The user is stuck inside its computer and drive around its filefolders. If he falls, he is teleported into the trash where he has to find the exit.
Technologies: C++, e-puck
School Project: Optimising the behavior of a flock of robots using PSO. Project in group of three.
Technologies: Java
Internship: Allow to easilly check and correct the models falses positives and negatives of the re-identification system.
Technologies: C++, OpenCv, Mqtt
Internship: Person tracking and re-identification system in a distributed multi camera environment.
Technologies: C++, Qt, OpenCv
Internship: Program to quickly extract and label training samples (from a video) for futur training. I developed this program during my research internship to automatize and assist me in this fastidious task (see code and instructions).
Technologies: C, Intel 8051
School Project: Creation and implementation of a communication protocol for a SPI Bus and deployment on an microcontroller within a mobile robot (project over 6 months, in a team of 8).
Technologies: C++, OpenGl
School Project: A 3d game where the gameplay is as subtle as the name.
Technologies: JavaScript, Dart, Json
Personal Project: Some astonishing extensions to make internet a better place.
Technologies: Java, Swing
School Project: MS Paint pales in comparison with this graphics editor developped with Java.
Technologies: Piano, Logic Studio
Personal Project: Musical composition for a short film in the context of an amateur film contest.
Technologies: C++, Qt, RegExp
Personal Project: I did develop this software to help me learn foreign vocabulary in a smarter way (it ask first the questions where the user often make mistake + other features).
Technologies: C++, Arduino
Personal Project: The prototype of our own google cars, which can follow a line on the ground and avoid obstacles. I developped it with friends and was in charge of the code part (demo).
Technologies: xHtml, CSS
School Project: A website made for an academic project. Still online here.
Technologies: C++, SFML
Personal Project: A 2D game engine developped in C++ with the SFML (Yes, I also did the graphics).
Technologies: C, SDL
Personal Project: A very simple scrolling shooter. It was my first "real" project, made during my high school years.
Don't hesitate to contact me if you want more information.