Full version of Persons Labeled contains 5711 images with annotated person instances (see download section). This repository is a sample with 595 images. Read more in our blog post.
Person segmentation is critical task in analysing humans on images for many real-world applications: action recognition, self-driving cars, video surveillance, mobile applications and much more.
We at Supervisely had our internal research on this field and we realized that there is a lack of data for this task. You can ask us: what about public datasets like COCO, Pascal, Mapillary and others? To answer this question i’ll better show you few examples from COCO dataset:
The quality of human segmentation in most public datasets is not satisfied our requirements and we had to create our own dataset with high quality annotations.
Sample (Supervisely format): tar archive (1 GB).
Full version (Supervisely format): tar archive (7,3 GB).
Sample project contains 595 images.
The full version contains 5711 images.
This Supervisely Person Dataset is made freely available to academic and non-academic entities for non-commercial purposes such as academic research, teaching, scientific publications, or personal experimentation.
The Supervisely Person Dataset data includes images obtained from the "pexels" website. Use of these images must respect the corresponding terms of use: