Warp-and-project tomography for rapidly deforming objects

Abstract

Computed tomography has emerged as the method of choice for scanning complex shapes as well as interior structures of stationary objects. Recent progress has also allowed the use of CT for analyzing deforming objects and dynamic phenomena, although the deformations have been constrained to be either slow or periodic motions. In this work we improve the tomographic reconstruction of time-varying geometries undergoing faster, non-periodic deformations. Our method uses a warp-and-project approach that allows us to introduce an essentially continuous time axis where consistency of the reconstructed shape with the projection images is enforced for the specific time and deformation state at which the image was captured. The method uses an efficient, time-adaptive solver that yields both the moving geometry as well as the deformation field. We validate our method with extensive experiments using both synthetic and real data from a range of different application scenarios.

TomoFluid: Reconstructing Dynamic Fluid From Sparse View Videos

“…A projection image from the 2,600 Re is shown in Figure 5(b). The normalized LII image, and reconstructed slices from our method (Ours) and Zang et al [53] method at 117°are shown respectively in Figure 5(c-e) indicating sharper results with reasonably finer soot structures that can be observed from our method. , where one soot raw image from the video captured by camera 1 is shown.…”

“…The appearance transfer based method (Okabe et al) for fluid modeling proposed by [36] is compared as the third baseline. The last baseline (Zang et al) reconstructs fast deforming object with internal structure by X-ray CT scanner via a so called warp-and-project strategy proposed in [53]. Since the problem we are tackling is highly underdetermined, improved results are achieved for each method by constraining the reconstruction to the visual hull [29] of each data.…”

“…, where one soot raw image from the video captured by camera 1 is shown. We compare normalized LII image taken at 117°(ground truth) with the same slice reconstructed by our method and Zang et al [53].…”

“…The sparse-view problem occurs also for dynamic X-ray reconstructions, since it is impossible to take a large number of projections representing the same state of the scanned object. Zang et al [52,53,50] proposed to jointly reconstruct the density and the deformation fields, by incorporating some spatial and temporal priors on these two fields. These approaches provide accurate reconstructions for the scanned objects.…”

“…This second approach allows the introduction of temporal priors that improve the reconstruction quality. However, reconstructing complex three-dimensional fluid flows using visible-light tomography remains a difficult task for several reasons: First, many phenomena like absorption, scattering, SART [18] [36] [53] Ours GT SART [18] [36] refraction, occlusion impact the captured data. Second, the involved setups are often sophisticated and require a difficult calibration step.…”