CATVis: Context-Aware Thought Visualization

TL;DR

This paper introduces CATVis, a novel framework that decodes visual representations from EEG signals to generate images, leveraging cross-modal alignment and re-ranking to improve semantic accuracy and image quality.

Contribution

It presents a new 5-stage EEG-to-image generation framework that outperforms existing methods in accuracy and image quality through innovative cross-modal techniques.

Findings

Outperforms SOTA in classification accuracy by 13.43%

Achieves 15.21% higher generation accuracy

Reduces Fréchet Inception Distance by 36.61%

Abstract

EEG-based brain-computer interfaces (BCIs) have shown promise in various applications, such as motor imagery and cognitive state monitoring. However, decoding visual representations from EEG signals remains a significant challenge due to their complex and noisy nature. We thus propose a novel 5-stage framework for decoding visual representations from EEG signals: (1) an EEG encoder for concept classification, (2) cross-modal alignment of EEG and text embeddings in CLIP feature space, (3) caption refinement via re-ranking, (4) weighted interpolation of concept and caption embeddings for richer semantics, and (5) image generation using a pre-trained Stable Diffusion model. We enable context-aware EEG-to-image generation through cross-modal alignment and re-ranking. Experimental results demonstrate that our method generates high-quality images aligned with visual stimuli, outperforming SOTA approaches by 13.43% in Classification Accuracy, 15.21% in Generation Accuracy and reducing Fr\'echet Inception Distance by 36.61%, indicating superior semantic alignment and image quality.