Novel 3D Scene Understanding Applications From Recurrence in a Single Image
This addresses the problem of efficient 3D scene analysis for computer vision applications, offering a novel approach that is competitive with methods using large datasets.
The paper tackled 3D scene understanding from a single image by discovering recurring patterns, achieving results as good as or better than existing supervised or unsupervised methods on a new benchmark with over 1,000 images.
We demonstrate the utility of recurring pattern discovery from a single image for spatial understanding of a 3D scene in terms of (1) vanishing point detection, (2) hypothesizing 3D translation symmetry and (3) counting the number of RP instances in the image. Furthermore, we illustrate the feasibility of leveraging RP discovery output to form a more precise, quantitative text description of the scene. Our quantitative evaluations on a new 1K+ Recurring Pattern (RP) benchmark with diverse variations show that visual perception of recurrence from one single view leads to scene understanding outcomes that are as good as or better than existing supervised methods and/or unsupervised methods that use millions of images.