Mahdi Mirhoseini

Dr. Mahdi Mirhoseini is a Postdoctoral Fellow at Université Laval specializing in deep learning, hyperspectral imaging, and high-dimensional data processing. He holds a Ph.D. in Electrical Engineering with extensive experience developing Vision Transformers, ConvLSTMs, and Attention-based networks. While his prior work focused on large-scale Earth observation (e.g., DESIS satellite imagery), the core computational challenge—extracting specific spectral signatures from massive, high-dimensional arrays—is identical to processing the 300-band VNIR data generated in biology labs. He is the lead author of multiple peer-reviewed publications utilizing hyperspectral data in journals such as IEEE JSTARS and the European Journal of Agronomy. In this project, Mahdi is responsible for standardizing the hyperspectral dataset and programming the open-source deconvolution models.

June 2026