arxiv: 2604.08063 · v1 · submitted 2026-04-09 · 💻 cs.CV

Recognition: 2 theorem links

· Lean Theorem

EEG2Vision: A Multimodal EEG-Based Framework for 2D Visual Reconstruction in Cognitive Neuroscience

Emanuele Balloni , Emanuele Frontoni , Chiara Matti , Marina Paolanti , Roberto Pierdicca , Emiliano Santarnecchi

Authors on Pith no claims yet

Pith reviewed 2026-05-10 17:48 UTC · model grok-4.3

classification 💻 cs.CV

keywords EEG-to-image reconstructiondiffusion modelsmultimodal large language modelslow-density EEGbrain-computer interfacesvisual decodingperceptual boosting

0 comments

The pith

A language-model-guided diffusion boost improves EEG image reconstructions even as electrode counts drop to 24.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper builds a modular EEG-to-image pipeline that first produces initial reconstructions by conditioning a diffusion model on brain signals recorded at varying resolutions, then applies a second stage to refine them. In that stage a multimodal large language model turns the rough output into a text description and an image-to-image diffusion process uses that description to sharpen geometry and appearance while trying to keep the content traceable to the original EEG. A sympathetic reader would care because high-channel EEG caps are expensive and stationary, so evidence that lower-channel portable devices can still yield acceptable images after this extra step suggests brain-to-image translation could move beyond the lab.

Core claim

Starting from EEG-conditioned diffusion reconstruction, the boosting stage uses a multimodal large language model to extract semantic descriptions and leverages image-to-image diffusion to refine geometry and perceptual coherence while preserving EEG-grounded structure, yielding consistent gains in perceptual metrics such as inception score and clear user preference even when channel count falls from 128 to 24.

What carries the argument

The prompt-guided post-reconstruction boosting mechanism that pairs multimodal LLM semantic extraction with image-to-image diffusion after the initial EEG-conditioned diffusion step.

Load-bearing premise

The language model must produce text descriptions that stay faithful to the EEG-derived content instead of introducing unrelated visual elements that the later diffusion step would then reinforce.

What would settle it

A blind user study or automated metric comparison in which the boosted low-channel images score worse than the unboosted versions on perceptual quality would show the boosting step fails to deliver its claimed improvement.

Figures

Figures reproduced from arXiv: 2604.08063 by Chiara Matti, Emanuele Balloni, Emanuele Frontoni, Emiliano Santarnecchi, Marina Paolanti, Roberto Pierdicca.

**Figure 2.** Figure 2: Qualitative results of the evaluation for different channel resolutions and [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗

**Figure 3.** Figure 3: Qualitative comparison of reconstructions across electrode configurations. [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗

**Figure 4.** Figure 4: Spatial distribution of decoding accuracy per electrode. (a) Top-1 Acc (b) Top-5 Acc [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗

read the original abstract

Reconstructing visual stimuli from non-invasive electroencephalography (EEG) remains challenging due to its low spatial resolution and high noise, particularly under realistic low-density electrode configurations. To address this, we present EEG2Vision, a modular, end-to-end EEG-to-image framework that systematically evaluates reconstruction performance across different EEG resolutions (128, 64, 32, and 24 channels) and enhances visual quality through a prompt-guided post-reconstruction boosting mechanism. Starting from EEG-conditioned diffusion reconstruction, the boosting stage uses a multimodal large language model to extract semantic descriptions and leverages image-to-image diffusion to refine geometry and perceptual coherence while preserving EEG-grounded structure. Our experiments show that semantic decoding accuracy degrades significantly with channel reduction (e.g., 50-way Top-1 Acc from 89% to 38%), while reconstruction quality slight decreases (e.g., FID from 76.77 to 80.51). The proposed boosting consistently improves perceptual metrics across all configurations, achieving up to 9.71% IS gains in low-channel settings. A user study confirms the clear perceptual preference for boosted reconstructions. The proposed approach significantly boosts the feasibility of real-time brain-2-image applications using low-resolution EEG devices, potentially unlocking this type of applications outside laboratory settings.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

EEG2Vision gives a workable modular pipeline for low-channel EEG to image but the 89-to-38 percent accuracy collapse shows the LLM booster is likely doing most of the semantic lifting once electrodes drop below 64.

read the letter

The paper's main contribution is a clear end-to-end setup that starts with EEG-conditioned diffusion, then uses a multimodal LLM to pull semantic prompts and runs image-to-image diffusion for refinement. They run the same pipeline on 128, 64, 32, and 24 channels and report both the semantic decoding hit and the perceptual metrics. That channel ablation plus the user study preference for boosted outputs is the part worth noting for anyone trying to move this kind of work out of the lab.

Referee Report

2 major / 2 minor

Summary. The paper introduces EEG2Vision, a modular framework for 2D visual reconstruction from EEG signals. It employs EEG-conditioned diffusion models for initial image generation and a boosting stage that uses a multimodal large language model to extract semantic descriptions from the initial reconstructions, followed by image-to-image diffusion to enhance perceptual quality. The authors evaluate the framework across varying EEG channel densities (128, 64, 32, 24 channels), reporting a significant drop in semantic decoding accuracy (e.g., 50-way Top-1 accuracy from 89% to 38%) but modest changes in FID (76.77 to 80.51), with the boosting providing up to 9.71% improvement in Inception Score and higher user preference. They conclude that this approach enhances the feasibility of real-time brain-to-image applications using low-resolution EEG devices outside laboratory settings.

Significance. Should the central claims be substantiated with detailed experimental protocols and evidence that the boosting stage preserves rather than supplants EEG-derived semantics, this work could have notable impact in advancing accessible brain-computer interface technologies for visual reconstruction. The systematic evaluation of channel reduction and the proposed boosting mechanism address practical challenges in EEG-based vision decoding, potentially bridging laboratory research with real-world applications. The inclusion of user studies adds to the perceptual validation, though the overall contribution depends on the robustness of the EEG grounding.

major comments (2)

[Abstract] Abstract: The claim that the boosting mechanism 'preserves EEG-grounded structure' is difficult to reconcile with the reported 50-way Top-1 accuracy collapse from 89% to 38% upon channel reduction. This performance drop indicates that low-density EEG (24-32 channels) supplies limited semantic content, suggesting that the multimodal LLM and subsequent diffusion steps may be the dominant contributors to the final image semantics rather than the EEG signal itself. This directly impacts the central claim regarding feasibility with low-resolution devices.
[Abstract / Experiments] Abstract / Experiments: No experimental details are provided regarding baselines, data splits, statistical significance tests, or the specific datasets used for the reported metrics (FID, IS, accuracy). Without these, it is not possible to assess whether the improvements from boosting are statistically meaningful or generalizable, which is load-bearing for the conclusion about real-time applications.

minor comments (2)

The abstract mentions 'up to 9.71% IS gains' but does not specify for which channel configuration this maximum is achieved or provide the full set of metrics for all configurations.
Consider adding more details on the multimodal LLM used (e.g., model name, prompting strategy) and the image-to-image diffusion parameters to improve reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and insightful comments on our manuscript. We appreciate the focus on clarifying the role of the boosting stage and ensuring experimental details are transparent. Below we provide point-by-point responses to the major comments. We have outlined revisions that will be incorporated in the next version to address the concerns raised.

read point-by-point responses

Referee: [Abstract] Abstract: The claim that the boosting mechanism 'preserves EEG-grounded structure' is difficult to reconcile with the reported 50-way Top-1 accuracy collapse from 89% to 38% upon channel reduction. This performance drop indicates that low-density EEG (24-32 channels) supplies limited semantic content, suggesting that the multimodal LLM and subsequent diffusion steps may be the dominant contributors to the final image semantics rather than the EEG signal itself. This directly impacts the central claim regarding feasibility with low-resolution devices.

Authors: We thank the referee for raising this critical point about the interplay between semantic decoding accuracy and the preservation claim. The 50-way Top-1 accuracy measures a separate EEG classification task on raw signals and does degrade with channel reduction, reflecting the inherent spatial limitations of low-density EEG. However, EEG2Vision generates the initial image via an EEG-conditioned diffusion model that directly conditions the latent space on EEG features, thereby embedding EEG-derived visual structure (such as coarse object layout and category cues) into the starting reconstruction. The boosting stage then performs image-to-image diffusion initialized from this EEG-generated image, with LLM descriptions extracted solely from the initial output; this process refines perceptual details and coherence without replacing the core EEG-grounded elements. The modest FID increase (76.77 to 80.51) across channel counts, combined with consistent IS gains from boosting (up to 9.71%) and user-study preference, indicates that sufficient EEG-derived information remains usable even at 24 channels. To strengthen this evidence, we will add an ablation experiment that compares boosted results when the image-to-image diffusion is initialized from the EEG reconstruction versus from noise or a generic image, quantifying how much of the final semantics trace back to the EEG input. This will directly support the feasibility claim for low-resolution devices. revision: yes
Referee: [Abstract / Experiments] Abstract / Experiments: No experimental details are provided regarding baselines, data splits, statistical significance tests, or the specific datasets used for the reported metrics (FID, IS, accuracy). Without these, it is not possible to assess whether the improvements from boosting are statistically meaningful or generalizable, which is load-bearing for the conclusion about real-time applications.

Authors: We agree that explicit experimental protocols are necessary to evaluate reproducibility, statistical significance, and generalizability. While the full manuscript reports the metrics and notes the use of a paired EEG-image dataset along with comparisons to prior methods, we acknowledge that the abstract and certain sections lack sufficient detail on data splits, baselines, and significance testing. In the revision we will expand the abstract with a concise statement of the dataset and evaluation protocol, add a dedicated 'Experimental Setup' subsection detailing the dataset (paired EEG recordings with visual stimuli), train/test splits (e.g., subject-independent partitioning), baseline methods (including non-boosted diffusion and prior EEG-to-image approaches), and statistical tests (paired t-tests for metric improvements and Wilcoxon tests for user preferences, with p-values reported). We will also include confidence intervals and significance markers in all result tables. These additions will allow readers to better judge the robustness of the boosting gains and the real-time application conclusions. revision: yes

Circularity Check

0 steps flagged

No significant circularity in empirical ML pipeline

full rationale

The paper presents an empirical ML framework for EEG-to-image reconstruction using diffusion models and LLM-based boosting. Performance is evaluated directly on standard metrics (Top-1 accuracy, FID, IS) across channel configurations with reported experimental results. No mathematical derivations, equations, fitted parameters presented as predictions, or load-bearing self-citations appear in the text. The central claims rest on observed performance numbers and user studies rather than any step that reduces to its own inputs by construction. This is the expected outcome for a self-contained empirical pipeline.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the empirical effectiveness of pre-trained diffusion models and multimodal LLMs when adapted to EEG conditioning and semantic prompting; no new physical entities or free parameters are introduced in the abstract.

axioms (1)

domain assumption Pre-trained diffusion models and multimodal LLMs can be conditioned on EEG signals and prompted to produce semantically useful refinements that preserve original structure
This assumption underpins the boosting mechanism described in the abstract.

pith-pipeline@v0.9.0 · 5546 in / 1304 out tokens · 58959 ms · 2026-05-10T17:48:47.488671+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

46 extracted references · 9 canonical work pages · 1 internal anchor

[1]

DreamDiffusion: Generating high- quality images from brain EEG signals.arXiv preprint arXiv:2306.16934, 2023

Bai, Y., Wang, X., Cao, Y.p., Ge, Y., Yuan, C., Shan, Y.: Dreamdiffusion: Gener- ating high-quality images from brain eeg signals. arXiv preprint arXiv:2306.16934 (2023)

work page arXiv 2023
[2]

Bisley, J.W., Goldberg, M.E.: Attention, intention, and priority in the parietal lobe33, 1–21.https://doi.org/https://doi.org/10.1146/ annurev-neuro-060909-152823,https://www.annualreviews.org/content/ journals/10.1146/annurev-neuro-060909-152823

work page doi:10.1146/annurev-neuro-060909-152823
[3]

IEEE Transactions on Multimedia26, 3430–3443 (2023)

Chen, B., Zhu, L., Zhu, H., Yang, W., Song, L., Wang, S.: Gap-closing matters: Perceptual quality evaluation and optimization of low-light image enhancement. IEEE Transactions on Multimedia26, 3430–3443 (2023)

2023
[4]

In: 2009 IEEE conference on computer vision and pattern recognition

Deng, J., Dong, W., Socher, R., Li, L.J., Li, K., Fei-Fei, L.: Imagenet: A large- scale hierarchical image database. In: 2009 IEEE conference on computer vision and pattern recognition. pp. 248–255. Ieee (2009)

2009
[5]

In: International Con- ference on Neural Computing for Advanced Applications

Deng, X., Bao, F., Liu, B., Li, Y., Zhang, L.: A study on image reconstruction based on decoding fmri through extracting image depth features. In: International Con- ference on Neural Computing for Advanced Applications. pp. 449–462. Springer (2024)

2024
[6]

Human Brain Map- ping15(2), 95–111 (2002).https://doi.org/https://doi.org/10.1002/hbm

Di Russo, F., Mart´ ınez, A., Sereno, M.I., Pitzalis, S., Hillyard, S.A.: Cortical sources of the early components of the visual evoked potential. Human Brain Map- ping15(2), 95–111 (2002).https://doi.org/https://doi.org/10.1002/hbm. 10010,https://onlinelibrary.wiley.com/doi/abs/10.1002/hbm.10010

work page doi:10.1002/hbm 2002
[7]

Nature Machine Intelligence pp

Doerig, A., Kietzmann, T.C., Allen, E., Wu, Y., Naselaris, T., Kay, K., Charest, I.: High-level visual representations in the human brain are aligned with large language models. Nature Machine Intelligence pp. 1–15 (2025)

2025
[8]

AI (2025)

Fares, A.: Understanding what the brain sees: Semantic recognition from eeg re- sponses to visual stimuli using transformer. AI (2025)

2025
[9]

Information Fusion p

Ferrante, M., Boccato, T., Rashkov, G., Toschi, N.: Towards neural foundation models for vision: Aligning eeg, meg, and fmri representations for decoding, en- coding, and modality conversion. Information Fusion p. 103650 (2025)

2025
[10]

The Llama 3 Herd of Models

Grattafiori, A., Dubey, A., Jauhri, A., Pandey, A., Kadian, A., Al-Dahle, A., Let- man, A., Mathur, A., Schelten, A., Vaughan, A., et al.: The llama 3 herd of models. arXiv preprint arXiv:2407.21783 (2024)

work page internal anchor Pith review Pith/arXiv arXiv 2024
[11]

Vision research41(10-11), 1409–1422 (2001)

Grill-Spector, K., Kourtzi, Z., Kanwisher, N.: The lateral occipital complex and its role in object recognition. Vision research41(10-11), 1409–1422 (2001)

2001
[12]

Scientific Reports14(1), 16436 (2024)

Guenther, S., Kosmyna, N., Maes, P.: Image classification and reconstruction from low-density eeg. Scientific Reports14(1), 16436 (2024)

2024
[13]

In: Proceedings of the Computer Vision and Pattern Recognition Conference

Guo, Z., Wu, J., Song, Y., Bu, J., Mai, W., Zheng, Q., Ouyang, W., Song, C.: Neuro-3d: Towards 3d visual decoding from eeg signals. In: Proceedings of the Computer Vision and Pattern Recognition Conference. pp. 23870–23880 (2025)

2025
[14]

Advances in neural information processing systems30(2017) 16 E

Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., Hochreiter, S.: Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems30(2017) 16 E. Balloni et al

2017
[15]

In: European Conference on Computer Vision

Huo, J., Wang, Y., Wang, Y., Qian, X., Li, C., Fu, Y., Feng, J.: Neuropictor: Refin- ing fmri-to-image reconstruction via multi-individual pretraining and multi-level modulation. In: European Conference on Computer Vision. pp. 56–73. Springer (2024)

2024
[16]

In: IJCAI

Jiao, Z., You, H., Yang, F., Li, X., Zhang, H., Shen, D.: Decoding eeg by visual- guided deep neural networks. In: IJCAI. vol. 28, pp. 1387–1393. Macao (2019)

2019
[17]

In: Proceedings of the 25th ACM interna- tional conference on Multimedia

Kavasidis, I., Palazzo, S., Spampinato, C., Giordano, D., Shah, M.: Brain2image: Converting brain signals into images. In: Proceedings of the 25th ACM interna- tional conference on Multimedia. pp. 1809–1817 (2017)

2017
[18]

Neural Computing and Applications34(8), 5979–5991 (2022)

Khare, S., Choubey, R.N., Amar, L., Udutalapalli, V.: Neurovision: perceived image regeneration using cprogan. Neural Computing and Applications34(8), 5979–5991 (2022)

2022
[19]

Seeing through the brain: image reconstruction of visual perception from human brain signals,

Lan, Y.T., Ren, K., Wang, Y., Zheng, W.L., Li, D., Lu, B.L., Qiu, L.: Seeing through the brain: image reconstruction of visual perception from human brain signals. arXiv preprint arXiv:2308.02510 (2023)

work page arXiv 2023
[20]

Journal of neural engineering15(5), 056013 (2018)

Lawhern, V.J., Solon, A.J., Waytowich, N.R., Gordon, S.M., Hung, C.P., Lance, B.J.: Eegnet: a compact convolutional neural network for eeg-based brain– computer interfaces. Journal of neural engineering15(5), 056013 (2018)

2018
[21]

Pattern Recognition100, 107085 (2020)

Li, D., Du, C., He, H.: Semi-supervised cross-modal image generation with gener- ative adversarial networks. Pattern Recognition100, 107085 (2020)

2020
[22]

Visual decoding and reconstruction via eeg embeddings with guided diffusion,

Li, D., Wei, C., Li, S., Zou, J., Qin, H., Liu, Q.: Visual decoding and reconstruction via eeg embeddings with guided diffusion. arXiv preprint arXiv:2403.07721 (2024)

work page arXiv 2024
[23]

In: ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Lopez, E., Sigillo, L., Colonnese, F., Panella, M., Comminiello, D.: Guess what i think: Streamlined eeg-to-image generation with latent diffusion models. In: ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). pp. 1–5. IEEE (2025)

2025
[24]

Journal of neural engineering18(4), 045013 (2021)

Lu, H.Y., Lorenc, E.S., Zhu, H., Kilmarx, J., Sulzer, J., Xie, C., Tobler, P.N., Watrous, A.J., Orsborn, A.L., Lewis-Peacock, J., et al.: Multi-scale neural decoding and analysis. Journal of neural engineering18(4), 045013 (2021)

2021
[25]

In: Proceedings of the 31st ACM international conference on multimedia

Lu, Y., Du, C., Zhou, Q., Wang, D., He, H.: Minddiffuser: Controlled image re- construction from human brain activity with semantic and structural diffusion. In: Proceedings of the 31st ACM international conference on multimedia. pp. 5899– 5908 (2023)

2023
[26]

Medical physics35(6Part1), 2541–2553 (2008)

Miao, J., Huo, D., Wilson, D.L.: Quantitative image quality evaluation of mr images using perceptual difference models. Medical physics35(6Part1), 2541–2553 (2008)

2008
[27]

In: 2022 International Conference on Intelligent Data Science Technologies and Applications (IDSTA)

Mishra, A., Raj, N., Bajwa, G.: Eeg-based image feature extraction for visual classification using deep learning. In: 2022 International Conference on Intelligent Data Science Technologies and Applications (IDSTA). pp. 181–188. IEEE (2022)

2022
[28]

Neural Computing and Applications 35(12), 9181–9192 (2023)

Mishra, R., Sharma, K., Jha, R.R., Bhavsar, A.: Neurogan: image reconstruction from eeg signals via an attention-based gan. Neural Computing and Applications 35(12), 9181–9192 (2023)

2023
[29]

arXiv preprint arXiv:2307.10246 , year=

Oota, S.R., Chen, Z., Gupta, M., Bapi, R.S., Jobard, G., Alexandre, F., Hinaut, X.: Deep neural networks and brain alignment: Brain encoding and decoding (survey). arXiv preprint arXiv:2307.10246 (2023)

work page arXiv 2023
[30]

European Journal of Neuroscience 61(1), e16636 (2025)

Ozkirli, A., Herzog, M.H., Jastrzebowska, M.A.: Computational complexity as a po- tential limitation on brain–behaviour mapping. European Journal of Neuroscience 61(1), e16636 (2025)

2025
[31]

IEEE Transactions on Pattern Analysis and Machine Intelligence 43(11), 3833–3849 (2020) EEG2Vision 17

Palazzo, S., Spampinato, C., Kavasidis, I., Giordano, D., Schmidt, J., Shah, M.: Decoding brain representations by multimodal learning of neural activity and vi- sual features. IEEE Transactions on Pattern Analysis and Machine Intelligence 43(11), 3833–3849 (2020) EEG2Vision 17

2020
[32]

https://huggingface.co/blog/stable_diffusion(2022), hugging Face Blog

Patil, S., Cuenca, P., Lambert, N., von Platen, P.: Stable diffusion with diffusers. https://huggingface.co/blog/stable_diffusion(2022), hugging Face Blog

2022
[33]

In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition

Prashnani, E., Cai, H., Mostofi, Y., Sen, P.: Pieapp: Perceptual image-error as- sessment through pairwise preference. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 1808–1817 (2018)

2018
[34]

Computer Methods and Programs in Biomedicine251, 108213 (2024)

Qian, D., Zeng, H., Cheng, W., Liu, Y., Bikki, T., Pan, J.: Neurodm: Decoding and visualizing human brain activity with eeg-guided diffusion model. Computer Methods and Programs in Biomedicine251, 108213 (2024)

2024
[35]

In: International conference on machine learning

Radford, A., Kim, J.W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al.: Learning transferable visual models from natural language supervision. In: International conference on machine learning. pp. 8748–8763. PmLR (2021)

2021
[36]

NeuroImage228, 117602 (2021)

Ren, Z., Li, J., Xue, X., Li, X., Yang, F., Jiao, Z., Gao, X.: Reconstructing seen im- age from brain activity by visually-guided cognitive representation and adversarial learning. NeuroImage228, 117602 (2021)

2021
[37]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

Rombach, R., Blattmann, A., Lorenz, D., Esser, P., Ommer, B.: High-resolution image synthesis with latent diffusion models. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 10684–10695 (2022)

2022
[38]

Advances in neural information processing systems29(2016)

Salimans, T., Goodfellow, I., Zaremba, W., Cheung, V., Radford, A., Chen, X.: Improved techniques for training gans. Advances in neural information processing systems29(2016)

2016
[39]

In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision

Singh, P., Dalal, D., Vashishtha, G., Miyapuram, K., Raman, S.: Learning ro- bust deep visual representations from eeg brain recordings. In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. pp. 7553–7562 (2024)

2024
[40]

In: ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Singh, P., Pandey, P., Miyapuram, K., Raman, S.: Eeg2image: image reconstruction from eeg brain signals. In: ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). pp. 1–5. IEEE (2023)

2023
[41]

Decod- ing natural images from eeg for object recognition,

Song, Y., Liu, B., Li, X., Shi, N., Wang, Y., Gao, X.: Decoding natural images from eeg for object recognition. arXiv preprint arXiv:2308.13234 (2023)

work page arXiv 2023
[42]

In: Proceedings of the IEEE conference on computer vision and pattern recognition

Spampinato, C., Palazzo, S., Kavasidis, I., Giordano, D., Souly, N., Shah, M.: Deep learning human mind for automated visual classification. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 6809–6817 (2017)

2017
[43]

1038/381520a0

Thorpe, S., Fize, D., Marlot, C.: Speed of processing in the human visual system 381(6582), 520–522.https://doi.org/10.1038/381520a0,https://doi.org/10. 1038/381520a0

work page doi:10.1038/381520a0
[44]

In: Proceedings of the 26th ACM international conference on Multimedia

Tirupattur, P., Rawat, Y.S., Spampinato, C., Shah, M.: Thoughtviz: Visualizing human thoughts using generative adversarial network. In: Proceedings of the 26th ACM international conference on Multimedia. pp. 950–958 (2018)

2018
[45]

In: Proceedings of the IEEE conference on computer vision and pattern recognition

Zhang, R., Isola, P., Efros, A.A., Shechtman, E., Wang, O.: The unreasonable effectiveness of deep features as a perceptual metric. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 586–595 (2018)

2018
[46]

In: International Conference on Neural Infor- mation Processing

Zheng, X., Cao, Z., Bai, Q.: An evoked potential-guided deep learning brain rep- resentation for visual classification. In: International Conference on Neural Infor- mation Processing. pp. 54–61. Springer (2020)

2020