Explore high-quality datasets for your AI and machine learning projects.
The NPM3D urban mobile mapping dataset includes instance labels and is intended for precise instance segmentation research on large‑scale LiDAR point clouds.
This is a SAR ship dataset derived from HRSID, primarily for ship detection and instance segmentation. It corrects annotation issues in the original dataset and removes large ships larger than 9,216 pixels to focus on small‑target detection.
MTPDSD is a comprehensive dataset containing 2,600 high‑quality pavement distress images, annotated with instance segmentation. It includes seven pavement distress categories: cracks, potholes, sealed cracks, patches, alligator cracks, facility covers, and expansion joints. In total, there are 3,596 annotated distress instances, each with a mask precisely aligned to the damage boundary.