IllusionVQA是一个包含光学幻觉和难以解释场景的数据集，旨在测试视觉语言模型在理解和软定位任务中的能力。GPT4V在理解任务中的准确率为62.99%，在定位任务中的准确率为49.7%，而人类在这两个任务中的准确率分别为91.03%和100%。数据集包含图像、问题、选项、答案、类别、推理和ID等特征，分为训练集和测试集，训练集包含4个样本，测试集包含1000个样本。数据集仅用于非商业研究目的

Hugging Face2024-03-26 更新130

Results from tests on differences in colour association parameters between high and low dimensional emotions.

情绪心理学

颜色感知

Results from tests on differences in colour association parameters between high and low dimensional emotions.

NIAID Data Ecosystem100

Tilt Illusion by Phase

视觉错觉

脑电图分析

# Overview This is the "Tilt Illusion" dataset. In brief, it contains EEG data for 36 subjects responding to the percieved orientation of a central target grating, that is titrated to appear vertic

OpenNeuro2024-03-09 更新150

Data and scripts from: Categorical perception of colour signals in a songbird

鸟类行为学

颜色感知

Categorical perception occurs when continuous variation in a stimulus (e.g., wavelength of color) is categorized by observers, with sharp changes in response occurring over a boundary and increased di

DataCite Commons2022-11-04 更新120

Luminance and Colorimetric Evaluation of ‘The Dress.’

资源简介：

相关数据集