Efficient Multi-view Surface Refinement with Adaptive Resolution Control
The existing stereo refinement methods optimize a surface representation using a multi-view photo-consistency functional. Such optimization is iterative and requires repeated computation of gradients over all surface regions, which is the bottleneck affec
- PDF / 6,842,160 Bytes
- 16 Pages / 439.37 x 666.142 pts Page_size
- 79 Downloads / 239 Views
stract. The existing stereo refinement methods optimize a surface representation using a multi-view photo-consistency functional. Such optimization is iterative and requires repeated computation of gradients over all surface regions, which is the bottleneck affecting adversely the computational efficiency of the refinement. In this paper, we present a flexible and efficient framework for mesh surface refinement in multi-view stereo. The newly proposed Adaptive Resolution Control (ARC) evaluates an optimal trade-off between the geometry accuracy and the performance via curve analysis. Then, it classifies the regions into the significant and insignificant ones using a graph-cut optimization. After that, each region is subdivided and simplified accordingly in the remaining refinement process, producing a triangular mesh in adaptive resolutions. Consequently, the ARC accelerates the stereo refinement by severalfold by culling out most insignificant regions, while still maintaining a similar level of geometry details that the state-of-the-art methods could achieve. We have implemented the ARC and demonstrated intensively on both public benchmarks and private datasets, which all confirm the effectiveness and the robustness of the ARC.
1
Introduction
Recovering a realistic 3D model from images is the ultimate goal of Multiple View Stereo (MVS) methods. Boosted by the public MVS benchmarks [7,15,16], the accuracy of stereovision has dramatically increased in last decade. It is believed the key factor to high accuracy is the final surface refinement step. With a triangular mesh representing the surface, refinement is a process of iterative adjustment of vertex locations by optimizing multi-view photo-consistency. Such iterative refinement is of heavy computation. The primary reason is the repeated computation of refinement gradient over all visible surface areas. Another reason is that mesh subdivision used in the refinement will dramatically increase the #vertices to be optimized. The higher density of mesh vertex also leads to slower mesh-related operations, e.g., mesh smoothing, visibility testing. Electronic supplementary material The online version of this chapter (doi:10. 1007/978-3-319-46448-0 21) contains supplementary material, which is available to authorized users. c Springer International Publishing AG 2016 B. Leibe et al. (Eds.): ECCV 2016, Part I, LNCS 9905, pp. 349–364, 2016. DOI: 10.1007/978-3-319-46448-0 21
350
S. Li et al.
(a) Initial noisy mesh
(b) ARC labeling
(c) Adaptive mesh density
(d) Final refined mesh
Fig. 1. With a noisy mesh as input (a), the ARC labels the mesh into two regions (b). Refinement applies only on the significant regions (orange), while the other insignificant regions (purple) will be culled out and simplified (c). This method greatly reduces the surface area to be optimized, but it is still able to produce valuable details (d). (Color figure online)
According to our observation, not all regions of refinement contribute equally to the geometry improvement. For example, most planar or low-textured reg
Data Loading...
