Research Highlight



Events and News

Mohammad Farazi’s WACV paper is Accepted, Congratulation!  This paper studies 3D dense shape correspondence, a key shape analysis application in computer vision and graphics. We introduce a novel hybrid geometric deep learning-based model that learns geometrically meaningful and discretization-independent features with a U-Net model as the primary node feature extraction module, followed by a successive spectral-based graph convolutional network. To create a diverse set of filters, we use anisotropic wavelet basis filters, being sensitive to both different directions and band-passes. This filter set overcomes the over-smoothing behavior of conventional graph neural networks. 

Yanshuai Tu and Xin Li’s STAR Protocols paper is Accepted, Congratulation! The hierarchical organization of the visual system results in topology, which often gets lost in the “raw” human retinotopic maps derived from BOLD fMRI recordings. Here, we present the protocol for generating topology-preserving and smooth retinotopic maps from human retinotopy fMRI data. We describe data pre-processing, 3D surface flattening, and selection of the region of interest (ROI), followed by smoothing of the retinotopic map within the ROI. This approach can be applied to visual cortex V1, V2 and V3 simultaneously.

Diffeomorphic Registration for Retinotopic Maps of Multiple Visual Regions

Yanshuai Tu’s BSF paper is Accepted, Congratulation! Retinotopic map, the mapping between visual inputs on the retina and neuronal responses on the cortical surface, is one of the central topics in vision science. Typically, human retinotopic maps are constructed by analyzing functional magnetic resonance responses to designed visual stimuli on the cortical surface. Although it is widely used in visual neuroscience, retinotopic maps are limited by the signal-to-noise ratio and spatial resolution of fMRI. One promising approach to improve the quality of retinotopic maps is to register individual subject’s retinotopic maps to a retinotopic template.