Moving Beyond Diffusion: Hierarchy-to-Hierarchy Autoregression for fMRI-to-Image Reconstruction
Researchers have developed MindHier, a novel framework for reconstructing images from fMRI data that moves beyond diffusion models. This new approach utilizes a scale-wise autoregressive method, incorporating a hierarchical fMRI encoder and a layer-wise alignment scheme with CLIP features. MindHier aims to mimic human visual perception by synthesizing global semantics before refining local details, resulting in faster inference times and more deterministic outputs compared to existing diffusion-based methods. AI
IMPACT Introduces a novel autoregressive framework for fMRI-to-image reconstruction, potentially improving brain-computer interfaces and neuroscience research.