metadata
dataset_info:
features:
- name: image
dtype: image
- name: label
dtype:
class_label:
names:
'0': confident
'1': unconfident
splits:
- name: train
num_bytes: 45225690.699
num_examples: 28707
- name: test
num_bytes: 11346824.156
num_examples: 7178
download_size: 53493485
dataset_size: 56572514.855000004
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
Facial Confidence Dataset
This dataset consists of 35,885 black and white images of peoples headshots. Each image is labeled "confident" or "unconfident" based on the facial expression being made. This dataset can primarily be used for image classification, specifically for behavioral analysis models.
References
This dataset was pulled from Kaggle and its original author is Sougat Ganguly. Link to Kaggle Dataset: https://www.kaggle.com/datasets/sougatganguly/confident-unconfident/data