Browsing by Subject "x-ray microtomography"

Fibrous meat analogues (FMAs) are an emerging category of synthetic plant-based food materials intended to mimic animal meat in both taste and mouthfeel. However, replicating the texture of meat with plant matter is far from trivial. Designing a manufacturing process that gives the produced FMA the desired properties requires proper characterisation of the material and its behaviour. One source of such knowledge could be obtained through computed x-ray absorption microtomography (µCT) --- a microscopic technique capable of producing a 3D model of the internal structure of the target object without requiring its disassembly. Unfortunately, while the scanning itself can be rather straightforward, issues arise when the recorded imagery is to be processed for analysis, as the researcher's choices of the employed methods, algorithms and their parameters may become a potential source of bias, affecting the end results in unintended ways. In this work, we examine the data processing pipeline of µCT studies on FMA samples, discuss the shortcomings of some of the employed methods, and attempt to address these problems by modifying the steps as needed, the aim being to reduce the reliance of the utilised methods on the researcher's judgment. Our main focus is in particular on the segmentation step, where the values provided by the µCT scan are interpreted and simplified to a form that subsequent algorithm-based analysis methods can operate on.