datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
EMID-Emotion-Matching
EMID-Emotion-Matching
orrzohar/EMID-Emotion-Matching is a derived dataset built on top of
the Emotionally paired Music and Image Dataset (EMID) from ECNU (ecnu-aigc/EMID).
It is designed for music ↔ image emotion matching with Qwen-Omni–style models.
Each example contains:
audio: mono waveform stored as datasets.Audio (HF Hub preview can play it)
sampling_rate: sampling rate used when decoding (typically 16 kHz)
image: a single image (datasets.Image)
same: bool, whether the audio… See the full description on the dataset page: https://huggingface.co/datasets/orrzohar/EMID-Emotion-Matching.child-emotion-drawings-pilot
Children's Emotional Drawings Pilot Dataset
A small balanced derived pilot dataset for experimental classification of emotional patterns in children's drawings.
Dataset
204 unique original drawings
492 total image instances
3 target classes: happiness, anxiety_depression, anger_aggression
Splits:
train: 432 images
validation: 30 images
test: 30 images
The dataset was created from the public anamelClassification dataset.
Original… See the full description on the dataset page: https://huggingface.co/datasets/stanislav-dykyi/child-emotion-drawings-pilot.Visual_Emotional_Analysisfacial_emotion_images
Facial Emotion Images
Dataset Summary
facial_emotion_images contains 15,109 grayscale facial images categorized into four distinct facial expressions: happy, neutral, sad, and surprise. It is structured for image classification and emotion recognition tasks.
Dataset Structure
Data Fields
image: PIL Image object containing the cropped face photo.
label: Class label corresponding to the emotion expression.
Class Labels &… See the full description on the dataset page: https://huggingface.co/datasets/gabriellidenor/facial_emotion_images.facial-emotion-recognition-datasetThe dataset consists of images capturing people displaying 7 distinct emotions
(anger, contempt, disgust, fear, happiness, sadness and surprise).
Each image in the dataset represents one of these specific emotions,
enabling researchers and machine learning practitioners to study and develop
models for emotion recognition and analysis.
The images encompass a diverse range of individuals, including different
genders, ethnicities, and age groups*. The dataset aims to provide
a comprehensive representation of human emotions, allowing for a wide range of
use cases.Dog_Emotion_Dataset_v2
Dataset Card for "Dog_Emotion_Dataset_v2"
The Dataset is based on a kaggle dataset
Label and its Meaning
0 : sad"
1 : angry"
2 : relaxed"
3 : happy"
Emotion_Video_Facial_Landmarks
Dataset Card for 478-Point Normalized 3D Facial Landmark Dataset
Dataset Description
This dataset provides pre-extracted, normalized 3D facial landmark features derived from the Video Emotion dataset. It is optimized for efficient training of emotion recognition and facial analysis models, bypassing the need to process large raw video files.
License: The extracted feature data in this CSV file is licensed under Apache 2.0. Note that the original source video files may… See the full description on the dataset page: https://huggingface.co/datasets/PSewmuthu/Emotion_Video_Facial_Landmarks.emotion_bias
Emotion Bias in Synthetic Face Generation
Description
This dataset accompanies the paper "Happy Young Women, Grumpy Old Men? Emotion Prompts as Demographic Selectors in AI Image Generation".
It contains 56,000 synthetic face images generated by eight state-of-the-art text-to-image (T2I) models across seven emotion prompt conditions, along with demographic attribute annotations (gender, race, age) and perceived attractiveness labels for each image.
The dataset is designed… See the full description on the dataset page: https://huggingface.co/datasets/mengtingwei/emotion_bias.emotion_detectionyolo-emotionsA merged emotions dataset was created using a highly curated subset of ExpW, FER2013 (enhanced with FER2013+), AffectNet (6 emotions), and RAF-DB in YOLO format, totaling approximately 155K samples. A YOLOv11-x model, fine-tuned on the WiderFace dataset for the bounding boxes, was used. The distribution is as follows:
TRAIN Set Class Distribution:
Class 0 (Angry): 8511 (6.84%)
Class 1 (Disgust): 6307 (5.07%)
Class 2 (Fear): 4249 (3.41%)
Class 3 (Happy): 37714 (30.30%)
Class 4… See the full description on the dataset page: https://huggingface.co/datasets/AdamCodd/yolo-emotions.Emotion_Video_Facial_Landmarks
Dataset Card for 478-Point Normalized 3D Facial Landmark Dataset
Dataset Description
This dataset provides pre-extracted, normalized 3D facial landmark features derived from the Video Emotion dataset. It is optimized for efficient training of emotion recognition and facial analysis models, bypassing the need to process large raw video files.
License: The extracted feature data in this CSV file is licensed under Apache 2.0. Note that the original source video files may… See the full description on the dataset page: https://huggingface.co/datasets/mac26/Emotion_Video_Facial_Landmarks.socratis_image_text_emotion
SOCRATIS: A benchmark of diverse open-ended emotional reactions to image-caption pairs.
ICCV WECIA Workshop 2023 (oral)
Project Page, Paper
We release a benchmark which contains 18K diverse emotions and reasons for feeling them on 2K image-caption pairs.
Our current preliminary findings have shown that Humans prefer human-written emotional reactions over machine-generated by more than two times.
We also find that current metrics fail to correlate with human preference… See the full description on the dataset page: https://huggingface.co/datasets/array/socratis_image_text_emotion.EMID-Emotion-Matching
EMID-Emotion-Matching
orrzohar/EMID-Emotion-Matching is a derived dataset built on top of
the Emotionally paired Music and Image Dataset (EMID) from ECNU (ecnu-aigc/EMID).
It is designed for music ↔ image emotion matching with Qwen-Omni–style models.
Each example contains:
audio: mono waveform stored as datasets.Audio (HF Hub preview can play it)
sampling_rate: sampling rate used when decoding (typically 16 kHz)
image: a single image (datasets.Image)
same: bool, whether the audio… See the full description on the dataset page: https://huggingface.co/datasets/lossminimilization/EMID-Emotion-Matching.karuna-emotion-dataset
Karuna Emotion Dataset
This dataset contains images annotated with the Navarasa emotion Karuna (sorrow).
Dataset Structure
Each sample includes:
image: facial image
navarasa: Karuna
intensity: low / medium / high
description: emotion-centric textual description based on facial affect cues
Creation Method
Image descriptions were generated using an open-source vision–language model.
Post-processing was applied to suppress non-emotional attributes and emphasize… See the full description on the dataset page: https://huggingface.co/datasets/Vaishnavi-29/karuna-emotion-dataset.