Welcome to my homepage !

I am a science enthusiast who enjoy discovering cultures and people all over the world. In my website you will find information about my research topics in computer science such as image processing or machine learning. You can also find some resources to the teachings I gave at University (C#, data mining, ...) and the scientific papers I published. Some of my hobbies are composing and playing music, traveling and taking pictures.

Kyoto workshop Manpu workshop

About me

Since October 2014, I am research assistant professor at Osaka Prefecture University in the intelligent media processing group which is supervised by Professor Koichi Kise. I was involved in a national CREST project and working with researchers from Tohoku, Kyushu and KEIO Universities until 2016. We use devices such as cameras, eye-trackers, EoGs or smart glasses in order to get information from the reading activity. From 2017 we started a new national CREST project with Todai and Kobe University about experiential supplements. The aim is to create new tools for sharing experiences, and especially for improving learning and education.

From 2012 to 2014, I was at the head of Research & Development at Gestform, a digitizing company of 340 people. I worked on many projects such as handwriting recognition, image classification, segmentation and recognition.

I am interested about signal analysis, image processing, machine learning, artificial intelligence and data mining. During my PhD.,I worked 3 years at LaBRI, the computer science laboratory of Bordeaux with Jean-Philippe Domenger and Nicholas Journet.

In 2009, I worked 6 months at INRIA about designing multitouch interfaces for 3D interactions in Pascal Guitton team. He is now director of INRIA research. In 2008, I worked 4 months at Kumamoto laboratory. My internship was about chaos theory and its application for random number generation. My advisor was Akio Tsuneda. You will find in my resume more information about my education, my experiences, my skills... I also have some recommendation letters from Pascal Guitton, Jean-Philippe Domenger and Akio Tsuneda.


Research and publications

The list of my publications is available on Google Scholar or the dblp website. I am doing research about image analysis, processing and understanding; machine learning and artificial intelligence; eye tracking and physiological sensors.

Reading analysis and mental state analysis with physiological sensors

eyeTrackers

Since the end of 2014, I am using different kinds of eye trackers for analysing the reading behavior at Osaka Prefecture University. The main idea is to create a reading-life log to quantify and qualify the daily life reading habits. I extended the research of a tool for counting the number of read words named the Wordometer. We also proposed several educational applications such as estimating the English skill (TOEIC score), the Japanese skill or the confidence while answering multiple choice questions.


Mangas analysis

My research group in Osaka Prefecture University is also working on mangas analysis. Analyzing drawings is a very challenging task due to their variety. We proposed a new method to analyze the story telling based on genre classification with a deep learning approach. I also proposed a method to generate dynamic image visualization based on eye tracking, automatic animation of characters based on deep learning, and preliminary research about physiological signal analysis (heartbeat, sweating, skin temperature, etc.) while reading mangas.

landmarks

Image analysis and machine learning

During my PhD at In Bordeaux university and Gestform digitizing company, I explored new techniques for image processing, analysis and comparison. I also studied machine learning, data mining and information retrieval. I used both image and text information for document image classification and indexation. I worked on many subjects such as : handwriting recognition based on Artificial Neural Network, document image classification based on bags of visual words, detection of check-box location and state, identity card robust and real-time retrieval, incremental learning with user interaction... I developed this techniques in C#, C++, R and applied them in industrial context.

Awards and grants

I was honored to win one best paper award at CIFED 2012.
In 2012, I got a young researcher grant to stay 3 weeks in the laboratory of my choice from LaBRI.
In 2013 I won the best student paper award at DRR 2013 for my works about semi-structured document image matching and recognition .
In 2015 I got Kakenhi young researcher grant (~4 millions JPY / ~30,000 EUR) to support my research about reading analysis and smart documents.
In 2017 I got the outstanding quality paper reviews award as reviewer of ICDAR 2017. I also got the outstanding service award for serving as local arrangement chair at ICDAR 2017.

Publication co-authors

Koichi Kise, Nicholas Journet, Charles Lima Sanches, Jean-Philippe Domenger, Kai Kunze, Motoi Iwata, Masahiko Inami, Tilman Dingler, Andreas Dengel, Benjamin Tag, Utpal Garain, Anne Vialard, and master students from Osaka Prefecture University.

Research supervised

  • Charles Lima Sanches (PhD): Reading behavior analysis via non-professional eye trackers to enhance knowledge acquisition.
  • Philippe Laroye (Master): Document image retrieval based on keypoint extraction
  • Anas Walaam (Master): Automatic analysis of checkbox in forms
  • Emma de Bont (Master): Text, logo and hanwritting segmentation in document images
  • Romain Husson (Master): Document image classification based on Bag of Visual Words
  • Étienne Bazin (Master): Handwritten digit recognition based on neural networks
  • Yann Boisset (Master): Document image analysis for predicting of OCR errors
  • Pauline Minary and Léa Galeron (Master): Automatic processing of receipt and expense reports
  • Nicholas Des Aunais (Master): Computer vision analysis for document duplicate retrieval
  • Ellina Guibert (Master): Combining text and image information for document image classification
  • Thomas Braud (Master): Analysing check marks on forms (OMR)
  • Fabien Michel (Master): Interactive and incremental learning for improving the classification of a image flow
  • Rémy Nashawati (Master): Improvement of handwritten digit recognition
  • Manuel Landsman (Master): Estimating reading skill of non-native readers with eye-tracker
  • Hitinui Robert (Master): Counting the number of read word with an EOG system
  • Clément Jacquet (Master): Creating a webplatform for logging the reading activity
  • Anne-Gaelle Schall (Master): Classifying reading and non-reading behaviors based on eye movement analysis
  • Marc Striker (Master): Detecting manga character faces with deep learning
  • Anastasia Garina (Master): Estimating the facial emotion of manga character with deep learning
  • Dominick Kratz (Master): Analysing the Japanese text difficulty for creating a recommender system to learn Japanese
  • David Pignato (Master): Space-Repetition system fro memorizing vocabulary
  • Kilian Brandt (Master): Generating animation of manga character based on deep learning
  • Hamza Ezzarouali El Boudri and Nabil Zaim (Master): Automatic text simplification system based on deep learning
  • Kevin Mounthanyvong (Master): Analysing the student understanding while reading a text
  • Mathilde Leconte (Master): Estimating the user interest in museums
  • Pierre Wobloski and Fabien Orea (Undergraduate): Creating a system for logging and sharing the web browsing experience
eyeGaze DRR2 icdarReview icdarChair
lena binarized lena binarized lena edges


Teachings

I am teaching data mining at Osaka Prefecture University, more specifically about the following topics:

  • Link analysis and PageRank (Google) algorithm.
  • TrustRank and combating web spam.
  • Frequent itemset mining and association rules.
  • Clustering.
  • Recommender systems: content-based and collaborative filtering.
  • Recommender systems: Latent Factor Models, and the Netflix Challenge.


I also taught computer science at Bordeaux University. You can find here some of courses, but they are all in French.

  • XHTML and CSS : 5 lessons.
  • Image processing
    • lesson 1 : Introduction, Histograms and binary operation, exercises and images used for the exercise
    • lesson 2 : Contour detection, image transformation, exercises and images used for the exercises. Another exercise about geometrical transformation and morpho-math
    • lesson 3 : Pattern recognition and segmentation and image analysis. Exercises
    • lesson 4 : Image synthesis. OpenGL introduction
  • C2i : 6 lessons - linux, internet, word, excel, open office and powerpoint.
  • UML : lesson, exercise and quick reference card.
  • C# : exercises.


Hobbies

Music composition

music

I like to play guitar (acoustic and electric) and piano. I enjoy jazz, blues and rock music. I use essentialy a Guibson SG with a Marshall amplifier. I also have a Wah-Wah and a Big Muff footswitches. You can find my first compositions on Soundcloud and the music I like on my itunes profile.


Traveling

For work or for leisure, I enjoy to visit many countries and cities to meet people all over the world such as in San Francisco, Kyoto, Siem Reap, Amsterdam, New York, Salzburg, Bagan, Venezia, Zaragosa... You can find more of my best shots on Instagram




Coding video games

When I was young I developed few video games with my teammates. Most of time we use C# and XNA or HTML5 and javascript. Here is some examples in HTML5: