Dataset Hub

The dataset contains multi‑view flash/no‑flash images together with corresponding camera poses and point‑cloud initializations generated by COLMAP.

The dataset evaluates algorithms' ability to align two Structure‑from‑Motion reconstructions under unknown relative pose and scale. It was built by recording multiple image sequences in a controlled indoor environment, using an ART‑2 tracking system to precisely track markers attached to the camera. A long sequence was processed with an offline incremental SfM pipeline to generate a scene point‑cloud and calibrate the transformation between the camera and tracker coordinate systems. Subsequently, twelve sequences were recorded in the tracked environment and processed with a keyframe‑based real‑time SLAM system.

Image dataset for testing OpenMVG, including 11 images of the Sceaux Castle, nine high‑resolution image datasets, an indoor planar dataset of 11 images, an outdoor school dataset of 4 images, and a Palm Desert micro‑drone dataset of 21 images captured with a DJI Mini2.

Guide3D is a dataset created by the University of Liverpool for 3D shape reconstruction of endovascular surgical tools. It comprises 8,746 high‑resolution biplane X‑ray video frames captured in realistic clinical settings and manually annotated. The dataset was acquired using a biplane X‑ray system with a torso vascular model and annotated with computer‑vision tools. Guide3D supports navigation and tool manipulation in endovascular surgery, aiming to improve segmentation and 3D reconstruction techniques for better surgical outcomes.