DATASET
Semantic image segmentation for training ML models to detect forms and body parts, accessories, tattoos and clothes. This dataset is perfect for setting atomisation for recommendation systems in online marketplaces and other e-Commerce projects.

People in underwear
and suits

E - commerce
Segmentation
splitting the image into semantically independent segments to determine the boundaries of objects.
Using machine learning in online trading
object annotation in the photo to train the system to recognize and interpret them.
Data labelling
object annotation in the photo to train the system to recognize and interpret them.
Computer Vision
7
types of labelling
6 787
images
9.41
Kaggle usability
Dataset's technical specifications
All labelling was done in PhotoShop
Images are provided in PNG and JPEG formats, masks are in PNG format
The names of the segmentation masks are consistent with the names of the original images
The image resolution is varied: from 200 to 2000 px
Dataset type's
Women I
1929 pairs “image - segmentation mask”
Semantic segmentation of both body parts and underwear for women
Labeled: Face and hair, Body above the waist and up to the forearms, Forearms, Hands, Legs, Tattoos, Smartphones
Women II
2468 pairs of “image - segmentation mask”
Detailed labelling of women in underwear. Special attention is paid to the labelling of the inner, outer and lower part of the breast in relation to the bra.
Labelled area: Face and hair, Inner chest, Outer chest chest, Lower chest, Hands, Forearms, Hands, Shoulders, Axillary area, Thighs, Lower Legs, Navel, Underwear top, Underwear bottom, Tattoos, Jewelry, Accessories and smartphones
Women III
184 pairs of “image - segmentation mask”
Semantic segmentation of underwear for women
Labeled: Face and hair, Swimsuit or underwear, Arms and body (above the waist), Legs, Jewelry, glasses and smartphones are highlighted in background color
Women IV
379 pairs of “image - segmentation mask”
Semantic segmentation of women in underwear with additional marking of the side press area.
Labelled area: Face and hair, Arms and neck, Forearms, Hips, Underwear, Stomach, Underwear top, Underwear bottom, Accessories and smartphones, additionally, the press area is labeled on top
Men I
28 pairs of “image - segmentation mask”
Detailed labelling of the body and clothing of men above the waist
Labeled area: Face and hair, Neck, Forearms and hands, Shoulders, Nipple area, Navel, Underwear, Tattoos, Smartphones and jewelry
Men II
309 pairs of “image - segmentation mask”
Semantic segmentation of men's body above the waist
Labeled area: Face and hair, Neck, Forearms and hands, Shoulders, Body from shoulders to waist, Smartphones
Men Neuro
1420 pairs of “image - segmentation mask”
Labelling created by the neural network for subsequent correction by the markup
Images are pre-labelled via neural network: Face and hair, Neck, Forearms and hands, Shoulders, Nipple area, Navel, Underwear, Tattoos, Smartphones and jewelry
Tell us about your project!
See other datasets