HERIDAL database contains over 68,750 image patches of wilderness acquired from an aerial perspective, 29,050 positive samples containing person as well as 39,700 negative samples .Of these, approximately 3,000 image patches are synthetically generated; the others are cropped from real images. Additionally, the HERIDAL database contains approximately 500 labelled, full-size 4,000 x 3,000 pixel real-world images. Full size images could be used for training (Fast or Faster R-CNN) as well as for testing purposes. At the moment 101 images have been selected for testing purposes.
A manuscript describing the dataset in detail: Dunja Božić-Štulić, Željko Marušić, Sven Gotovac: Deep Learning Approach on Aerial Imagery in Supporting Land Search and Rescue Missions, International Journal of Computer Vision, 2019.
Labeled
Patches
Citations and Data Usage policy
This collection is freely available to browse, download, and use for commercial, scientific and educational purposes as outlined in the Creative Commons Attribution 3.0 Unported License. For additional details questions may be directed to dgotovac@fesb.hr.
Please be sure to include the following citations in your work if you use this data set:
Dunja Božić-Štulić, Željko Marušić, Sven Gotovac: Deep Learning Approach on Aerial Imagery in Supporting Land Search and Rescue Missions, International Journal of Computer Vision, 2019.