Versatile Framework with Semantic and Structural guidance for Image Reconstruction from Brain Activity
Researchers have developed a novel two-stage framework called MindDiffuser for reconstructing images from brain activity. This method first uses CLIP text embeddings with Stable Diffusion to generate an image based on semantic content. In the second stage, it refines the image by aligning structural information using decoded visual features from CLIP, improving accuracy and controllability. AI
IMPACT Enhances brain-computer interface capabilities by improving image reconstruction accuracy and controllability from neural data.