Scaling the U-net: segmentation of biodegradable bone implants in high-resolution synchrotron radiation microtomograms

Abstract

Highly accurate segmentation of large 3D volumes is a demanding task. Challenging applications like the segmentation of synchrotron radiation microtomograms (SRμCT) at high-resolution, which suffer from low contrast, high spatial variability and measurement artifacts, readily exceed the capacities of conventional segmentation methods, including the manual segmentation by human experts. The quantitative characterization of the osseointegration and spatio-temporal biodegradation process of bone implants requires reliable, and very precise segmentation. We investigated the scaling of 2D U-net for high resolution grayscale volumes by three crucial model hyper-parameters (i.e., the model width, depth, and input size). To leverage the 3D information of high-resolution SRμCT, common three axes prediction fusing is extended, investigating the effect of adding more than three axes prediction. In a systematic evaluation we compare the performance of scaling the U-net by intersection over union (IoU) and quantitative measurements of osseointegration and degradation parameters. Overall, we observe that a compound scaling of the U-net and multi-axes prediction fusing with soft voting yields the highest IoU for the class “degradation layer”. Finally, the quantitative analysis showed that the parameters calculated with model segmentation deviated less from the high quality results than those obtained by a semi-automatic segmentation method.

Degradation behavior and osseointegration of Mg–Zn–Ca screws in different bone regions of growing sheep: a pilot study

“…The effective pixel size was 1.06 µm, which was binned during reconstruction to a pixel size of 3.18 µm. Stitched tomograms were reconstructed using a reconstruction framework implemented in MATLAB [ 18 , 19 ] and employing the ASTRA toolbox for tomographic back projection [ 20 , 21 ]. The reconstructed data sets were filtered using an iterative nonlocal means filter [ 22 ].…”

Multiscale morphological analysis of bone microarchitecture around Mg-10Gd implants

“…The processing of the SRμCT data have been described by Küger et al [ 30 ]. Prior to segmentation of the lacunar and vascular porosity, first, the bone tissue and the implant were automatically segmented using a U-Net convolutional neural network [ 56 , 57 ]. Employing the segmentation method developed by Nunez et al [ 58 ], the bone porosities (lacunar and vascular porosity) were extracted from the bone label.…”

Synchrotron microtomography reveals insights into the degradation kinetics of bio-degradable coronary magnesium scaffolds

“…However, such segmentation can be challenging and time consuming because of the complexity of the corrosion process resulting in complex corrosion fronts and corrosion phases with material densities close to the initial metallic phase. Recently, convolutional networks were successfully used for 3D image segmentation of various materials, including degradable magnesium bone implants [ [17] , [18] , [19] , [20] ].…”