The reconstruction of 3D scenes from a pair of stereo images/videos that were captured from slightly displaced camera view points presents an important computer vision task for a variety of applications, including human motion capture and novel view synthesis, robotics, autonomously moving cars and drones, or 3D city reconstruction. In the context of the Internet of Things, networks of smart cameras can capture and exchange information on the observed 3D scene content. A key problem in stereo analysis is the determination of dense point correspondences in the input stereo pair, in order to generate a disparity (or depth) map as illustrated in Figure 1. High resolution images acquired by modern cameras impose special difficulties for stereo matching such as computational complexity and matching ambiguities, affecting both run-time and 3D reconstruction quality. In our work, we investigate hierarchical stereo matching to resolve the aforementioned difficulties by using image pyramids and simple but effective disparity propagation techniques.
M. Nezveda, M. Gelautz: "3D Scene Reconstruction via Hierarchical Stereo Matching"; Talk: Vienna young Scientists Symposium, Wien; 06-09-2016 - 06-10-2016; in: "Vienna Young Scientists Symposium", (2016), ISBN: 978-3-9504017-2-1; 72 - 73.
Click into the text area and press Ctrl+A/Ctrl+C or ⌘+A/⌘+C to copy the BibTeX into your clipboard… or download the BibTeX.