Projects

Selected public projects by Melvin Sevi in generative modeling, computer vision, and applied mathematics.

Selected projects from my MVA studies and earlier work. My current research interests are described on the About page, and published research is listed under Publications.

Generation diversity with IP-Adapter

An MVA project exploring diversity in image generation with IP-Adapter.

Code

Visual perception with text-to-image diffusion models

An MVA project studying the use of text-to-image diffusion models for referring image segmentation.

Code · Report

Expectation-maximization for hidden Markov models

An implementation of the expectation-maximization algorithm for parameter estimation in hidden Markov models.

Code

Motif detection in graphs

An earlier project on detecting motifs in graphs.

Code

Explore more on GitHub.