Dataset

Human photos data collection

7 344

Images
272

People
19.4

GB data
  • 7 344
    Images
  • 272
    Unique faces
  • 19.4
    GB data
  • Re-identification
  • Computer Vision
  • Classification
  • Machine Learning
Task type
  • Data Collection
The dataset consists of various people's photos. Each person provided one selfie portrait and 26 photos from everyday life. Portrait is a selfie made with the frontal smartphone camera, a person is captured alone in a "passport style" format. Photos from everyday life include shots taken in different locations, with different people, but a person from the Selfie Portrait is always present on all of them.

Everyday life photos have different formats. There are photos captured with the phone, group photos, holiday photos. A person in the photos can be moving or expressing emotions. However, he or she is present in all the photos and his/her face can be easily recognized.

Selfie portrait, on the contrary, has strict criteria:
  • the photo is cropped above shoulders,
  • the person is depicted alone,
  • there are no sunglasses or a medical mask on the face,
  • calm expression, no smiling

Most of the respondents are from Southeast Asia and Africa. But there are also respondents from Latin America, Eastern Europe and the USA.
The average participant’s age is 27 years with a standard deviation of ± 7 years.

Men make up 60% of the total participants in this dataset.

Our company customizes datasets upon request – collects additional data and conducts high-quality markup for your project.
TOP by Kaggle:
May 2023
TOP by Kaggle:
May 2023
607
Views
62
Downloads
10.0
Usability (max 10)
Views
607
Downloads
62
Usability (max 10)
10.00
Error get alias