I had the great pleasure to collaborate on the image analysis side of a very interesting projet in biology.
This article, entitled “A Vascular Endothelial Growth Factor-Dependent Sprouting Angiogenesis Assay Based on an In Vitro Human Blood Vessel Model for the Study of Anti-Angiogenic Drugs” is available here :
We have just released a demo of a system for art style recognition in paintings (using only the image and no metadata).
The demo is accessible here.
I am now visiting the RIKEN AIP center as a visiting researcher for a month.
I have been selected among the best reviewers at NIPS 2015 :
NIPS 2015 Awards
Lots of changes have happened recently.
The biggest new being that I am now an associate professor at Université Paris-Dauphine.
Also, the following papers have been accepted/published since my last update :
- “Averaging Covariance Matrices for EEG Signal Classification based on the CSP: an Empirical Study”, FY, Fabien Lotte and Masashi Sugiyama presented in EUSIPCO 2015
- “Importance-weighted covariance estimation for robust common spatial pattern”, Alessandro Balzi, FY and Masashi Sugiyama accepted for publication in Pattern Recognition Letters
- “Geometry-Aware Principal Component Analysis for Symmetric Positive Definite Matrices”, Inbal Horev, FY and Mavashi Sugiyama, accepted to ACML.
Two technical report are also available on arxiv :
- Supervised LogEuclidean Metric Learning for Symmetric Positive Definite Matrices
- Challenge IEEE-ISBI/TCB : Application of Covariance matrices and wavelet marginals
The publication page of this site should be updated accordingly soon.
… my affiliation has changed since last October …
As the lab moved from Tokyo Institute of Technology to the University of Tokyo, I am now a postdoc from UTokyo.
As published by the organizers of this challenge, one of my approach has been ranked first on the Bone Texture Characterization Challenge organized for the last ISBI conference.
This approach used some Wavelet Marginals (described and used in Wavelet Kernel Learning for building kernels). Two other approaches. based on covariance matrices, have been less succesfull and seem to have overfitted.
I have just joined Masashi Sugiyama’s machine learning laboratory as a postdoctoral researcher.
Shortly before my phd defense, I submitted a paper to MLSP 2013 (Southampton).
This paper, entitled “A Review Of Kernels On Covariance Matrices For Bci Applications “, has been accepted. The draft will be available soon after the conference.
I defended my Phd thesis one month ago (June 19th) at Rouen university.
The jury was composed of (in alphabetical order) :
The manuscipt (in french) will be available soon.