Datasets:

1saccc
/

ASD-NT-child-dataset

text string
================================================================================
Qwen3-VL-30B-A3B CSV+视频情感分析结果
================================================================================

分析CSV文件: /media/mount_sda/chensuncan/ASD_NT_child/ASD/AD001/AD001_01_开心_小孩吃棒棒糖_clean.csv
分析视频文件: /media/mount_sda/chensuncan/ASD_NT_child/ASD/AD001/AD001_01_开心_小孩吃棒棒糖_face.mp4
原始总帧数: 106
处理后总帧数: 106
使用模型: Qwen3-VL-30B-A3B

--------------------------------------------------------------------------------
数据预处理信息:
--------------------------------------------------------------------------------
删除的列: face_id, confidence, success
删除的全0 AU列数: 21
剩余列数: 30
剩余列名: frame, timestamp, gaze_0_x, gaze_0_y, gaze_0_z, gaze_1_x, gaze_1_y, gaze_1_z, gaze_angle_x, gaze_angle_y, pose_Tx, pose_Ty, pose_Tz, pose_Rx, pose_Ry, pose_Rz, AU01_r, AU04_r, AU05_r, AU07_r, AU17_r, AU20_r, AU25_r, AU01_c, AU04_c, AU05_c, AU07_c, AU17_c, AU20_c, AU25_c

================================================================================
模型分析结果:
================================================================================

[Behavioral Caption]:
The participant exhibits a consistent pattern of facial muscle activity and gaze behavior throughout the video. Initially, there is a brief period of subtle brow elevation and eye widening, followed by sustained activation of the chin raiser and upper lip raiser, indicating a gradual engagement of the facial musculatur...

[Final Emotion Label]:
happy

[Russell Circumplex Output]:
[240, 180, 'Pleasure is inferred as high due to the sustained activation of the lip corner puller and chin raiser, which are associated with positive affect. Arousal is inferred as high due to the consistent activation of the upper lid raiser and the presence of gaze shifts, indicating heightened attention and engageme...

【代码计算的Russell环曼哈顿距离分析】
模型推断的pleasure值: 240.0
模型推断的arousal值: 180.0

所有情绪类别的曼哈顿距离（从近到远排序）:
Happy: \|240.0-(272)\| + \|180.0-(178)\| = 34.0
Delighted: \|240.0-(192)\| + \|180.0-(159)\| = 69.0
Glad: \|240.0-(228)\| + \|180.0-(85)\| = 107.0
Excited: \|240.0-(203)\| + \|180.0-(258)\| = 115.0
Pleased: \|240.0-(253)\| + \|180.0-(64)\| = 129.0
Aroused: \|240.0-(80)\| + \|180.0-(192)\| = 172.0
Content: \|240.0-(262)\| + \|180.0-(-40)\| = 242.0
Satisfied: \|240.0-(261)\| + \|180.0-(-57)\| = 258.0
Astonished: \|240.0-(55)\| + \|180.0-(300)\| = 305.0
At ease: \|240.0-(195)\| + \|180.0-(-88)\| = 313.0

最小曼哈顿距离匹配: Happy
最小曼哈顿距离: 34.0
匹配坐标: pleasure=272, arousal=178


================================================================================
分析完成!
================================================================================
Qwen3-VL-30B-A3B CSV+视频情感分析结果
================================================================================

分析CSV文件: /media/mount_sda/chensuncan/ASD_NT_child/ASD/AD001/AD001_02_开心_小孩过生日_clean.csv
分析视频文件: /media/mount_sda/chensuncan/ASD_NT_child/ASD/AD001/AD001_02_开心_小孩过生日_face.mp4
原始总帧数: 573
处理后总帧数: 573
使用模型: Qwen3-VL-30B-A3B

--------------------------------------------------------------------------------
数据预处理信息:
--------------------------------------------------------------------------------
删除的列: face_id, confidence, success
删除的全0 AU列数: 12
剩余列数: 39
剩余列名: frame, timestamp, gaze_0_x, gaze_0_y, gaze_0_z, gaze_1_x, gaze_1_y, gaze_1_z, gaze_angle_x, gaze_angle_y, pose_Tx, pose_Ty, pose_Tz, pose_Rx, pose_Ry, pose_Rz, AU01_r, AU02_r, AU04_r, AU07_r, AU09_r, AU15_r, AU17_r, AU20_r, AU25_r, AU26_r, AU45_r, AU01_c, AU02_c, AU04_c, AU07_c, AU09_c, AU15_c, AU17_c, AU20_c, AU...

================================================================================
模型分析结果:
================================================================================

[Behavioral Caption]: The participant exhibits a dynamic sequence of facial expressions and gaze shifts throughout the video. Initially, the participant displays a neutral expression with minimal facial muscle activity. As the video progresses, the participant's gaze shifts frequently, indicating engagement with the vi...

[Final Emotion Label]: happy

[Russell Circumplex Output]:[272, 178, 'The participant's facial expressions show a consistent activation of the inner and outer brow raisers, indicating heightened attention. The mouth corners are lifted, suggesting a positive emotional response. The intensity of these facial movements varies, with periods of increase...

【代码计算的Russell环曼哈顿距离分析】
模型推断的pleasure值: 272.0
模型推断的arousal值: 178.0

所有情绪类别的曼哈顿距离（从近到远排序）:
Happy: \|272.0-(272)\| + \|178.0-(178)\| = 0.0
Delighted: \|272.0-(192)\| + \|178.0-(159)\| = 99.0
Pleased: \|272.0-(253)\| + \|178.0-(64)\| = 133.0
Glad: \|272.0-(228)\| + \|178.0-(85)\| = 137.0
Excited: \|272.0-(203)\| + \|178.0-(258)\| = 149.0
Aroused: \|272.0-(80)\| + \|178.0-(192)\| = 206.0
Content: \|272.0-(262)\| + \|178.0-(-40)\| = 228.0
Satisfied: \|272.0-(261)\| + \|178.0-(-57)\| = 246.0
Astonished: \|272.0-(55)\| + \|178.0-(300)\| = 339.0
At ease: \|272.0-(195)\| + \|178.0-(-88)\| = 343.0

最小曼哈顿距离匹配: Happy
最小曼哈顿距离: 0.0

End of preview. Expand in Data Studio

ASD-NT儿童情绪表达数据集

这是一个包含自闭症(ASD)和神经典型(NT)儿童情绪表达视频的数据集。

数据集结构

ASD-NT child/
├── ASD/          # 自闭症儿童视频
│   ├── AD001/
│   ├── AD002/
│   └── ...
└── NT/           # 神经典型儿童视频
    ├── NT001/
    ├── NT002/
    └── ...

文件命名格式

视频文件: {ID}_{片段号}_{情绪}_{场景}_face.mp4
CSV文件: {ID}_{片段号}_{情绪}_{场景}_clean.csv

情绪类别

恐惧 (停电、出现蛇)
悲伤 (小孩哭)
愤怒 (铲葡萄)

使用说明

本数据集仅供研究使用。

Downloads last month: 49