Recovering 3D scenes from 2D images is an under-constrained task; optimal estimation depends upon knowledge of the underlying scene statistics.
In this project we introduce the Southampton-York Natural Scenes dataset (SYNS: https://syns.soton.ac.uk), which provides comprehensive scene statistics useful for understanding biological vision and for improving machine vision systems. In order to capture the diversity of environments that humans encounter, scenes were surveyed at random locations within 25 indoor and outdoor categories.
Each survey includes:
- Spherical LiDAR range data
- High-dynamic range spherical imagery and
- A panorama of stereo image pairs.
We have made the dataset public and free of charge and envisage many uses. We have already used the dataset to derive statistics on surface attitude, conditioned on scene category and viewing elevation, and we are now employing the database to train deep network classifiers to distinguish depth edges from other kinds of edges.
To find out more:
- Adams, W.J., Elder, J.H., Graf, E.W., Leyland, J., Lugtigheid, A.J. & Muryy, A. (2016). The Southampton-York Natural Scenes (SYNS) dataset: Statistics of surface attitude. Scientific Reports 6. (16 pages)
- Adams, W.J. & Elder, J.H. (2014). Effects of specular highlights on perceived surface convexity. PLOS Computational Biology 10(5): e1003576. doi:10.1371/journal.pcbi.1003576.