Biopolymer Networks: Image Analysis, Reconstruction and Modeling

Abstract

The aim of this work is to extract the architecture of biopolymer networks from 3D images. This is motivated to further understand non-affine regimes, found during network formation and in low-density biopolymer networks, where the geometry of a network has a fundamental role in defining its mechanical properties.Firstly, developed image analysis tools were extended to 3D and contributed tohigh-performance open-source libraries for image analysis. These developments in isotropic wavelets will help in extracting realistic networks by removing spurious noise generated during image acquisition.Secondly, images of biopolymer gels from transmission electron microscopy (TEM), were used to reliably extract the network architecture. The imaged material was also studied with small-angle x-ray scattering (SAXS), and the comparison showed a strong agreement for network-size features.Thirdly, spatial graphs were extracted from the image. A one-to-one map is provided between image and graph, keeping all the geometric information from the image. This then opened the door to using analytical tools from the complex networks field to characterize images.Finally, statistical distributions extracted from three graph properties were used to reconstruct a completely in-silico network using a simulated annealing technique to generate new networks.This can be used as a computational exploration tool of how network behavior depends on network architecture.

Structural and hemodynamic properties of murine pulmonary arterial networks under hypoxia-induced pulmonary hypertension

“…The skeletonization S ( F , D ) refers to a thinned version of the foreground voxel complex F , which preserves the connection and branching pattern of F (Figure 1(b)). We compute S using Couprie and Bertrand’s “Asymmetric Thinning” algorithm, which iteratively removes voxels from F following the framework of “critical kernels.” 39–41 This creates a thinned copy of F , in which each branch is one-voxel wide. During the “Asymmetric Thinning” algorithm, a choice must periodically be made to chose what voxels, within a group of voxels X ⊆ F , should be included in S .…”

“…By choosing this voxel, the resulting skeletonized network S is centered in F (Figure 1(b)). 40 We refer to S as the “skeletonization” or the “centerline network” of F .…”

Automatic quantification framework to detect cracks in teeth

“…For the quantification of cracked teeth, we used the Isotropic Wavelet package[20] developed in the Insight Toolkit (ITK)[21]. ITK is an open-source, cross-platform system that provides developers with an extensive suite of software tools for image analysis.…”

“…ITK is an open-source, cross-platform system that provides developers with an extensive suite of software tools for image analysis. The ITKIsotropicWavelets module is available as an optional external module for ITK since version ITK-4.13, and can also be downloaded from the development site[20]. Following the same practices as other ITK functions and external modules, it provides Python wrappings for the C++ code to facilitate its use.…”

“…The package ITKIsotropicWavelets implements a Steerable Wavelet Pyramid [20] in the frequency domain. The wavelet pyramid (figure 3) represents a multi-resolution analysis (MRA) framework, providing information at multiple scales, from small details to bigger features.…”