AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Cross-Subject Mind Decoding from Inaccurate Representations

Yangyang Xu, Bangzhen Liu, Wenqi Shao, Yong Du, Shengfeng He, Tingting Zhu

ICCV 2025 · Image generation › Autoregressive and latent generation · approach: Latent diffusion with learned tokenizers · 1 citations (OpenAlex)

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 changedWhereMeasuredEffect
learn bidirectional fMRI-representation mapping instead of unidirectional mapping (ablation)BAI cross-subject decoder · NSD (avg of 4 subjects)PixCorr: .281 -> .318Improved
add Subject Bias Modulation Module (SBMM) (ablation)BAI cross-subject decoder · NSD (avg of 4 subjects)AlexNet(2): 83.7% -> 97.3%Improved
add SRM and VCM instead of direct addition of predicted representations (ablation)decoding stage with ControlNet · NSD (avg of 4 subjects)PixCorr: .228 -> .318Improved
add Semantic Refinement Module to Direct Addition (ablation)Direct Addition · NSD (avg of 4 subjects)CLIP: 89.8% -> 95.7%Improved
add Visual Coherence Module to Direct Addition (ablation)Direct Addition · NSD (avg of 4 subjects)PixCorr: .228 -> .298Improved
use single cross-subject model instead of subject-specific models (ablation)BAI · NSD (avg of 4 subjects)PixCorr: .321 -> .318No clear effect
adapt only SBMM on new subject instead of training from scratch (500 samples) (ablation)BAI · NSD Subj07, 500 samplesPixCorr: .083 -> .145Improved
increase number of new-subject samples from 500 to 1,500 (ablation)BAI (adapted) · NSD Subj07PixCorr: .145 -> .152Improved
compare with MindBridge adaptation at 1,500 samplesBAI (adapted) · NSD Subj07, 1,500 samplesAlexNet(5): 92.6% (MindBridge) vs 92.3%Mixed
compare with NeuroPictor on SSIMBAI · NSD (avg of 4 subjects)SSIM: .375 (NeuroPictor) vs .356Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy