Human Part Segmentation Datasets

Datasets

We introduced two new robotics datasets for human semantic part segmentation: Freiburg Sitting People and Freiburg People in Disaster. They provide high resolution data for experiments on ground and aerial robot segmentation applications.


• Freiburg Sitting people constitutes a dataset with two hundred and one images of six different people in multiple viewpoints and in a wide range of orientations. The data is segmented into 14 parts. DATA
  • We additionally provide the training set two people training and four for testing. 2Train4Testing



• Freiburg People in Disaster dataset is a collection of images captured using an aerial platform. The dataset consists of images and corresponding segmentation masks in an environment that mimics disaster scenario, with clutter and heavy occlusion around. For this specific dataset only a set of images are publicly available with its correspondent segmentation mask. DATA