Automated thresholding and analysis of microCT scanned bread dough

作者：Jensen S; Samanta S; Chakrabarti Bell S; Regenauer Lieb K; Siddique K H M; Wang S^*

来源：Journal of Microscopy, 2014, 256(2): 100-110.

DOI：10.1111/jmi.12163

• Bread dough microstructure
• visualization systems and software
• visualization techniques and methodologies
• x-ray microtomography

Computer aided x-ray microtomography is an increasingly popular method to investigate the structure of materials. Continuing improvements in the technique are resulting in increasingly larger data sets. The analysis of these data sets generally involves executing a static workflow for multiple samples and is generally performed manually by researchers. Executing these processes requires a significant time investment. A workflow which is able to automate the activities of the user would be useful. In this work, we have developed an automated workflow for the analysis of microtomography scanned bread dough data sets averaging 5GB in size. Comparing the automated workflow with the manual workflow indicates a significant amount of the time spent (33% in the case of bread dough) on user interactions in manual method. Both workflows return similar results for porosity and pore frequency distribution. Finally, by implementing an automated workflow, users save the time which would be required to manually execute the workflow. This time can be spent on more productive tasks. %26lt;br%26gt;Lay Description Technological advances in x-ray scanning techniques have resulted in larger, more complex microtomographic datasets. Processing these datasets can be both a time consuming and oftentimes repetitive task as datasets of similar materials tend to have similar characteristics. What if there was a better way to analyze these datasets? Our research has investigated using computer programming languages instead of researchers to automatically perform tasks involved in the analysis of microtomographic datasets of ten scanned bread dough samples. Our results highlighted the benefits of using computers to automate the analysis process, demonstrating that nearly 33% of the time required is due to researchers interacting with the analysis programs and not from the analysis itself. The quantitative data provided by the automated workflow are nearly identical to results found by researchers. This research highlights the benefits of using automated computer processing for analysis of microtomographic datasets, which would allow researchers to spend their efforts elsewhere.

• 出版日期2014-11
• 单位CSIRO