Research map / Image generation / Autoregressive and latent generation
Autoregressive Visual Decoding from EEG Signals
Sicheng Dai, Hongwang Xiao, Shan Yu, Qiwei Ye
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use pre-trained LaBraM encoder instead of ATM (ablation) | AVDE (VAR generator) · THINGS-EEG reconstruction (all-subject avg, ablation) | SSIM: 0.351 -> 0.366 | Improved |
| use pre-trained LaBraM encoder instead of EEGNet (ablation) | AVDE (VAR generator) · THINGS-EEG reconstruction (all-subject avg, ablation) | SSIM: 0.323 -> 0.366 | Improved |
| use pre-trained LaBraM encoder instead of NICE (ablation) | AVDE (VAR generator) · THINGS-EEG reconstruction (all-subject avg, ablation) | SwAV: 0.613 -> 0.586 | Improved |
| use VAR next-scale generator instead of unCLIP pipeline (ablation) | AVDE (LaBraM encoder) · THINGS-EEG reconstruction (all-subject avg, ablation) | SSIM: 0.346 -> 0.366 | Improved |
| use VAR instead of LDM-4 diffusion (ablation) | AVDE (LaBraM encoder) · THINGS-EEG reconstruction (all-subject avg, ablation) | SwAV: 0.609 -> 0.586 | Improved |
| use VAR instead of DiT-XL/2 diffusion (ablation) | AVDE (LaBraM encoder) · THINGS-EEG reconstruction (all-subject avg, ablation) | SwAV: 0.594 -> 0.586 | Improved |
| use pre-trained LaBraM encoder instead of ATM (ablation) | AVDE (VAR generator) · THINGS-EEG reconstruction (all-subject avg, ablation) | PixCorr: 0.141 -> 0.147 | Improved |
| replace multi-stage unCLIP SDXL pipeline with autoregressive VAR generator | EEG-to-image decoder · Inference efficiency (A100, batch 1) | inference time (ms): 310.4 -> 91.2 | Improved |
| AVDE vs ATM in cross-subject retrieval on Subject-07 (AVDE below baselines) | AVDE · THINGS-EEG 200-way retrieval (cross-subject) | top-1 accuracy: 0.074 vs NICE 0.102 | Mixed |
| evaluate AVDE on EEG-ImageNet vs ATM | AVDE · EEG-ImageNet 80-way classification | top-1 accuracy: Sub-06 0.187 -> 0.329 | Improved |
Closest papers by meaning
- Visual Decoding and Reconstruction via EEG Embeddings with Guided Diffusion (NeurIPS 2024 · similarity 0.91)
- MindAlign: Bridging EEG, Vision, and Language for Zero-Shot Visual Decoding (NeurIPS 2026 · similarity 0.91)
- EEG2Video: Towards Decoding Dynamic Visual Perception from EEG Signals (NeurIPS 2024 · similarity 0.91)
- What Does the Brain See? Multiview Neural Representation to Demystify the Brain-Visual Alignment (NeurIPS 2026 · similarity 0.91)
- Brain decoding: toward real-time reconstruction of visual perception (ICLR 2024 · similarity 0.90)
- Decoding Natural Images from EEG for Object Recognition (ICLR 2024 · similarity 0.90)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
