Title: Hyperspectral Reconstruction of Skin Through Fusion of Scattering Transform Features. 

Abstract: Hyperspectral Imaging (HSI) is an established technique with an array of applications, but its use is limited due to both practical and technical issues associated with spectral devices. In this talk, I will discuss work my collaborators and I have done as part of the ICASSP 2024 Grand Challenge on Hyperspectral Skin Vision on reconstructing HSI of human skin from matching RGB images and an infrared band. To address this problem, we proposed a model based on the matching of features from the scattering transform - a type of convolutional neural network with predefined wavelet filters. I will introduce this model, show our results, and discuss possible improvements to the model.