KAITANG2003/cat-emotion-dataset

这是一个用于猫面部情绪识别的综合数据集，包含13,619张标记图像，涵盖4种情绪类别。每张图像都标注了四种情绪之一，并包含视觉特征描述和思维链分析。数据集结构包括图像文件名、视觉特征详细描述(耳朵位置、眼睛状态、身体姿势等)、情绪分类的思维链分析以及最终情绪标签(攻击性、快乐、放松或悲伤)。标签分布显示：攻击性5.2%(707张)、快乐29.6%(4,035张)、放松61.7%(8,399张)、悲伤3.5%(478张)。该数据集是通过合并和标准化三个独立的猫情绪数据集创建的。

A comprehensive dataset for cat facial emotion recognition containing 13,619 labeled images across 4 emotion categories. Each image is labeled with one of four emotions and includes visual feature descriptions and chain-of-thought analysis. The dataset structure includes image filename, detailed description of visual features (ears position, eye state, body posture, etc.), chain-of-thought analysis explaining the emotion classification, and final emotion label (Aggressive, Happy, Relaxed, or Sad). Label distribution: Aggressive 5.2% (707 images), Happy 29.6% (4,035 images), Relaxed 61.7% (8,399 images), Sad 3.5% (478 images). This dataset was created by merging and normalizing three separate cat emotion datasets.

搜集汇总

背景与挑战

背景概述

该数据集是一个多模态猫情感识别数据集，包含13,592个样本，涵盖Relaxed、Happy、Aggressive和Sad四种情感类别，其中Relaxed情感占比最高（61.7%）。每个样本提供猫图像、视觉特征描述、思维链分析和情感标签，适用于情感识别研究。