Sign inGet started
Dataset · Hugging Face

illusion-brain/omnibrain-4d

OmniBrain 4D — THINGS-EEG 四轴完整视图 把 THINGS-EEG 重组成四轴完整保留的结构:image × repetition × channel × time, 外加派生的频率轴。 供需要试次结构、时间进程、频带分解的工作直接使用。 不是把 EEG 压平成 (n_trial, n_channel) 的二维表。 ⚠️ 先读这一条:train 与 test 的轴结构完全不同 test/ train/ 形状 (200, 80, 63, 250) (16,540, 4, 63, 250) 图像数 200 16,540 重复数 80 4 每被试试次 16,000 66,160 10 被试总计 160,000 661,600 试次间信度 +0.072 +0.005 磁盘 4.5 GB 20.2 GB 这是 THINGS-EEG 的原始设计,不是错误: 训练集:图像多(16,540)、重复少(4) → 用于学跨图像泛化… See the full description on the dataset page: https://huggingface.co/datasets/illusion-brain/omnibrain-4d.

Author / organizationillusion-brain
License (as declared)cc-by-4.0
Downloads0 (as reported by Hugging Face)
Likes1
Created18 October 2026
Last updated at source18 October 2026
eegneurosciencethings-eegbrainvisiontensor
Embed this entry
<iframe src="https://silytics.si/embed/item/cb57560f-2d7c-4e7f-b408-4be2dfda30dc" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Recent changes

  • 18 Oct 2026 License: none → cc-by-4.0
  • 18 Oct 2026 Description: none → OmniBrain 4D — THINGS-EEG 四轴完整视图 把 THINGS-EEG 重组成四轴完整保留的结构:image × repetition × channel × time, 外加派生的频率轴。 供需要试次结构、时间进程、频带分解的工作直接使用。 不是把 EEG 压平成 (n_trial, n_channel) 的二维表。 ⚠️ 先读这一条:train 与 test 的轴结构完全不同 test/ train/ 形状 (200, 80, 63, 250) (16
  • 18 Oct 2026 Topics added task_categories:feature-extraction, task_categories:image-classification, eeg, neuroscience, things-eeg, brain, vision, tensor

Changes detected by Silytics between refreshes of the source.

More from illusion-brain

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.