Title: Neuro-3D: Towards 3D Visual Decoding from EEG Sginals

• Ref: https://arxiv.org/abs/2411.12248
• Summary: To understand how the brain perceives and processes 3D visual stimuli in the real world, this paper introduced a new neuroscience task: decoding 3D visual perception from EEG signals, a neuroimaging technique that enables real-time monitoring of neural dynamics enriched with complex visual cues. The researchers developed the following models:

1. EEG-3D: An open-source multimodal EEG dataset from 12 subjects who viewed 72 categories of 3D objects, including both videos and images.

2. Neuro-3D: A diffusion-based colored point cloud decoder—a 3D visual decoding framework based on EEG signals that recovers both the shape and color of 3D objects.