Color Vision Test Dataset
收藏资源简介:
该数据集名为Color Vision Test Dataset,由中国科学院大学、中国科学院、浙江实验室、浙江大学的研究人员创建,旨在评估大型视觉语言模型在颜色视觉方面的能力。数据集包含了5450个颜色视觉测试图像,涵盖了数字、动物、形状、字母或字符、物体五大类,并分为容易和困难两个任务级别。数据集的创建过程严格,采用了Daltonlens工具包生成图像,并通过Coblis工具评估图像质量。数据集可用于评估和比较不同大型视觉语言模型在颜色视觉测试中的有效性,为安全敏感领域中的应用提供了重要依据。
The dataset, named Color Vision Test Dataset, was created by researchers from the University of Chinese Academy of Sciences, Chinese Academy of Sciences, Zhejiang Lab, and Zhejiang University. It is designed to evaluate the color vision capabilities of large vision-language models. The dataset contains 5,450 color vision test images covering five categories: numbers, animals, shapes, letters/characters, and objects, and is divided into two task difficulty levels: easy and hard. The dataset was rigorously constructed, with images generated using the Daltonlens toolkit and image quality assessed via the Coblis tool. This dataset can be used to evaluate and compare the effectiveness of different large vision-language models in color vision tests, providing an important basis for applications in safety-sensitive fields.

- 1Assessing Color Vision Test in Large Vision-language Models中国科学院大学, 中国科学院, 浙江实验室, 浙江大学 · 2025年



