EG 2016 - Short Papers

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/15066

Browse

Browsing EG 2016 - Short Papers by Subject "I.4.1 [Image Processing and Computer Vision]"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Multi-Focus Plenoptic Simulator and Lens Pattern Mixing for Dense Depth Map Estimation
    (The Eurographics Association, 2016) Ferreira, Rodrigo; Cunha, Joel; Goncalves, Nuno; T. Bashford-Rogers and L. P. Santos
    Light field cameras capture a scene's multi-directional light field with one image, allowing the estimation of depth. In this paper, we introduce a fully automatic method for depth estimation from a single plenoptic image running a RANSAC-like algorithm for feature matching. The novelty about our method is the use of different focal-length lenses for multiple depth map refining, generating a dense depth map for future all-in-focus renders. We also present a plenoptic simulator which produces a plenoptic dataset from a 3D computer rendered scene. This simulator, which is unique, as far as we known, allows testing of plenoptic oriented algorithms since it can reproduce datasets with desired scene characteristics, providing the depth ground truth for error measurement. This work is a on-going project with promising results.
  • Thumbnail Image
    Item
    A Simple and Effective Method to Detect Orthogonal Vanishing Points in Uncalibrated Images of Man-Made Environments
    (The Eurographics Association, 2016) Simon, Gilles; Fond, Antoine; Berger, Marie-Odile; T. Bashford-Rogers and L. P. Santos
    This paper presents an effective and easy-to-implement algorithm to compute orthogonal vanishing points in uncalibrated images of man-made scenes. The main contribution is to estimate the zenith and the horizon line before detecting the vanishing points, using simple properties of the central projection and exploiting accumulations of oriented segments around the horizon. Our method is fast and yields an accuracy comparable, and even better in some cases, to that of state-of-the-art algorithms.