ETD system

Electronic theses and dissertations repository

 

Tesi etd-01102017-094240


Thesis type
Tesi di dottorato di ricerca
Author
BALDACCI, ANDREA
email address
baldacci85@hotmail.it
URN
etd-01102017-094240
Title
Novel Methods and Algorithms for Presenting 3D Scenes
Settore scientifico disciplinare
ING-INF/05
Corso di studi
INGEGNERIA
Commissione
tutor Prof. Avvenuti, Marco
relatore Dott. Ganovelli, Fabio
Parole chiave
  • computer animation
  • multimedia content production
  • video similarity
  • 2D vector field comparison
  • image based reconstruction
  • image based modeling
  • surface reconstruction
  • depth maps fusion
  • energy minimization on the GPU
  • stereoscopy
  • image based rendering
Data inizio appello
14/02/2017;
Consultabilità
completa
Riassunto analitico
In recent years, improvements in the acquisition and creation of 3D models gave rise to<br>an increasing availability of 3D content and to a widening of the audience such content<br>is created for, which brought into focus the need for effective ways to visualize and<br>interact with it.<br>Until recently, the task of virtual inspection of a 3D object or navigation inside a 3D<br>scene was carried out by using human machine interaction (HMI) metaphors controlled<br>through mouse and keyboard events.<br>However, this interaction approach may be cumbersome for the general audience.<br>Furthermore, the inception and spread of touch-based mobile devices, such as smartphones<br>and tablets, redefined the interaction problem entirely, since neither mouse nor<br>keyboards are available anymore. The problem is made even worse by the fact that these<br>devices are typically lower power if compared to desktop machines, while high-quality<br>rendering is a computationally intensive task.<br>In this thesis, we present a series of novel methods for the easy presentation of 3D<br>content both when it is already available in a digitized form and when it must be acquired<br>from the real world by image-based techniques. In the first case, we propose<br>a method which takes as input the 3D scene of interest and an example video, and it<br>automatically produces a video of the input scene that resembles the given video example.<br>In other words, our algorithm allows the user to replicate an existing video, for<br>example, a video created by a professional animator, on a different 3D scene.<br>In the context of image-based techniques, exploiting the inherent spatial organization<br>of photographs taken for the 3D reconstruction of a scene, we propose an intuitive<br>interface for the smooth stereoscopic navigation of the acquired scene providing an immersive<br>experience without the need of a complete 3D reconstruction.<br>Finally, we propose an interactive framework for improving low-quality 3D reconstructions<br>obtained through image-based reconstruction algorithms. Using few strokes on<br>the input images, the user can specify high-level geometric hints to improve incomplete<br>or noisy reconstructions which are caused by various quite common conditions<br>often arising for objects such as buildings, streets and numerous other human-made<br>functional elements.
File