Estimation of the multivariate self-similarity parameters for operator fractional Brownian motions. Published in TSP.
Paper Toolbox More detailsAutomatic selection of the Discrete Mumford-Shah functional hyperparameter for parameter-free image reconstruction preserving contours. Published in SIVP.
Paper Toolbox More detailsI am currently a Postdoc in the Department of Mathematics at San Diego State University, California, USA. I am working on the 2D empirical wavelet transform under the supervision Jérôme Gilles. Empirical wavelets consist of adaptative wavelet filterbanks built from a partitioning of the Fourier domain of the data under study. I am developing wavelet frames defined on detected Fourier supports from any mother wavelet using diffeomorphisms.
- Multivariate self-similarity: scaling exponents, wavelet transform, bootstrapping, statistical hypothesis testing, spectral clustering, high dimension.
- Image denoising and contour detection: Mumford-Shah model, hyperparameter tuning.