57786_2 - Geometry Super-Resolution by Example Thales...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Geometry Super-Resolution by Example Thales Vieira Alex Bordignon Thomas Lewiner Matm´ ıdia Laboratory Department of Mathematics, PUC–Rio. Rio de Janeiro, Brazil www.matmidia.mat.puc-rio.br/ { thalesv,alexlaier,tomlew } Luiz Velho Visgraf Laboratory IMPA. Rio de Janeiro, Brazil www.impa.br/˜lvelho Figure 1. Super-resolution of a scanned keyboard with a high-resolution scan of one of its keys: original model (top), reconstructed with the detail exemplar of a key scanned at high resolution (middle) and reconstructed with the exemplar key generated by low resolution accumulation (bottom). Abstract —The acquisition of high-resolution 3D models still requires delicate and time-consuming processes. In particular, each detail of the object should be scanned separately, although they may be similar. This can be simplified by copying a small set of details at different places of the model, synthesizing high geometric resolution from details exemplars, as introduced in this paper for three different contexts : when the detail exemplars are scanned separately at high resolution, when they are synthesized or edited from other models, or when they are obtained by accumulating repeated instances of the detail in the low-resolution scan. The main challenge here is to correctly register the high-resolution details with the low resolution model. To address this issue, this work proposes a careful resolution manipulation of 3D scans at each step of an automatic registration pipeline, combined with a robust selection of alignments. This results in a fully automatic process for geometry super-resolution by example. Experiments on syn- thetic and real data sets show applicability in different contexts, including resolution increase, noise removal by example and geometric texture insertion. Keywords-Super-resolution; 3D Registration; 3D Scanning; Surface Reconstruction; I. INTRODUCTION With the advent of cheaper scanning devices, 3D model acquisition became more popular. However, the scanning of high-resolution model remains a complex and time- consuming task. Regardless of the equipment used, the 3D scanning hardware imposes a limit in the highest achievable resolution of a single scan. Additionally, the acquired data is contaminated by sensor noise, further compromising the effective model resolution. There is an inherent trade-off between scanning time and model resolution: While for digitizing an object in low resolution only a few broad scans may suffice ; To scan the same object in high resolution many close scans would be needed to capture each fine detail, even if they are repeated in the model. This trade-off can be improved if we reuse a single exemplar of each fine detail, turning the scan process simpler, faster and still producing models with high resolution. This goal can be accomplished using super- resolution by example ! Figure 2. Intelligent geometry insertion: substituting the exemplar detail for a synthetic texture, a new object can be obtained by super-resolution,...
View Full Document

Page1 / 8

57786_2 - Geometry Super-Resolution by Example Thales...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online