HBID24K: A New Benchmark Dataset for Vulnerable Houbara Bustard and Intruder Detection

Houbara Bustard and Intruders Dataset for the manuscript with the title "HBID24K: A New Benchmark Dataset for Vulnerable Houbara Bustard and Intruder Detection in Wildlife Monitoring"It contains 2 major folders HBID24K folder and Codes Folder.HBID24K Folder contains totals 24,318 .JPG files and 24,318 .json files in 4 subfolders. These 4 subfolders are train_images, validation_images, train_labels, and validation_labels.Codes folder contains the benchmarking codes for 10 state-of-the-art object detection techniques, Faster R-CNN, Cascade R-CNN, RetinaNet, FCOS, RepPoints, ATSS, Deformable DETR, Sparse R-CNN, YOLOv10, and DETR.<br>Project/├── HBID24K/│ ├── train_images/ (19,454 .JPG files)│ ├── train_labels/ (19,454 .json files)│ ├── validation_images/ (4,864 .JPG files)│ └── validation_labels/ (4,864 .json files)│├── Codes/│ ├── Faster_R-CNN/│ ├── Cascade_R-CNN/│ ├── RetinaNet/│ ├── FCOS/│ ├── RepPoints/│ ├── ATSS/│ ├── Deformable_DETR/│ ├── Sparse_R-CNN/│ ├── YOLOv10/│ └── DETR/The dataset on Figshare has now been organized into 5 smaller parts, so that it is easier for the user to download. We also added HBID24_Samples.zip which contains only 2027 images with annotations, in case the user doesn’t want to download the complete dataset.<br>