A Large Scale Homography Benchmark
This provides a standardized benchmark for researchers in computer vision working on homography estimation, though it is incremental as it builds on existing datasets like 1DSfM.
The authors tackled the problem of robust homography estimation by introducing Pi3D, a large-scale dataset of 1000 planes in 10,000 images, and HEB, a benchmark with 226,260 homographies and 4M correspondences, enabling rigorous evaluation of methods and establishing state-of-the-art results.
We present a large-scale dataset of Planes in 3D, Pi3D, of roughly 1000 planes observed in 10 000 images from the 1DSfM dataset, and HEB, a large-scale homography estimation benchmark leveraging Pi3D. The applications of the Pi3D dataset are diverse, e.g. training or evaluating monocular depth, surface normal estimation and image matching algorithms. The HEB dataset consists of 226 260 homographies and includes roughly 4M correspondences. The homographies link images that often undergo significant viewpoint and illumination changes. As applications of HEB, we perform a rigorous evaluation of a wide range of robust estimators and deep learning-based correspondence filtering methods, establishing the current state-of-the-art in robust homography estimation. We also evaluate the uncertainty of the SIFT orientations and scales w.r.t. the ground truth coming from the underlying homographies and provide codes for comparing uncertainty of custom detectors. The dataset is available at \url{https://github.com/danini/homography-benchmark}.