A machine learning based approach to the segmentation of micro CT data in archaeological and evolutionary sciences

O'Mahoney, Thomas; Mcknight, Lidija; Lowe, Tristan; Mednikova, Maria B.; Dunn, Jacob C.

doi:10.1101/859983

Cited by 3 publications

(3 citation statements)

References 77 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Machine learning segmentation techniques have been demonstrated to result in fewer misclassified pixels than both Otsu and watershed methods, and to perform well on noisy images ( Andrew 2018 ; Dunmore et al. 2018 ; O'Mahoney et al. 2019 ).…”

Section: Methodsmentioning

confidence: 99%

A Shrewd Inspection of Vertebral Regionalization in Large Shrews (Soricidae: Crocidurinae)

Smith

Angielczyk

2022

Integrative Organismal Biology

View full text Add to dashboard Cite

The regionalization of the mammalian spinal column is an important evolutionary, developmental, and functional hallmark of the clade. Vertebral column regions are usually defined using transitions in external bone morphology, such as the presence of transverse foraminae or rib facets, or measurements of vertebral shape. Yet the internal structure of vertebrae, specifically the trabecular (spongy) bone, plays an important role in vertebral function, and is subject to the same variety of selective, functional, and developmental influences as external bone morphology. Here we investigated regionalization of external and trabecular bone morphology in the vertebral column of a group of shrews (family Soricidae). The primary goals of this study were to: 1) determine if vertebral trabecular bone morphology is regionalized in large shrews, and if so, in what configuration relative to external morphology; 2) assess correlations between trabecular bone regionalization and functional or developmental influences; and 3) determine if external and trabecular bone regionalization patterns provide clues about the function of the highly modified spinal column of the hero shrew Scutisorex. Trabecular bone is regionalized along the soricid vertebral column, but the configuration of trabecular bone regions does not match that of the external vertebral morphology, and is less consistent across individuals and species. The cervical region has the most distinct and consistent trabecular bone morphology, with dense trabeculae indicative of the ability to withstand forces in a variety of directions. Scutisorex exhibits an additional external morphology region compared to unmodified shrews, but this region does not correspond to a change in trabecular architecture. Although trabecular bone architecture is regionalized along the soricid vertebral column, and this regionalization is potentially related to bone functional adaptation, there are likely aspects of vertebral functional regionalization that are not detectable using trabecular bone morphology. For example, the external morphology of the Scutisorex lumbar spine shows signs of an extra functional region that is not apparent in trabecular bone analyses. It is possible that body size and locomotor mode affect the degree to which function is manifest in trabecular bone, and broader study across mammalian size and ecology is warranted to understand the relationship between trabecular bone morphology and other measures of vertebral function such as intervertebral range of motion.

show abstract

Section: Methodsmentioning

confidence: 99%

A Shrewd Inspection of Vertebral Regionalization in Large Shrews (Soricidae: Crocidurinae)

Smith

Angielczyk

2022

Integrative Organismal Biology

View full text Add to dashboard Cite

show abstract

“…For example, the method in [ 16 ] did not distinguish between soft tissue, bones, and teeth within the mummy. Low resolution volumetric scans of mummies were considered in [ 17 , 18 ], with segmentation performed one slice at a time using classical machine learning techniques. In both works, ridged artifacts and loss of detail were evident in the results obtained.…”

Section: Introductionmentioning

confidence: 99%

Automated segmentation of microtomography imaging of Egyptian mummies

et al. 2021

View full text Add to dashboard Cite

Propagation Phase Contrast Synchrotron Microtomography (PPC-SRμCT) is the gold standard for non-invasive and non-destructive access to internal structures of archaeological remains. In this analysis, the virtual specimen needs to be segmented to separate different parts or materials, a process that normally requires considerable human effort. In the Automated SEgmentation of Microtomography Imaging (ASEMI) project, we developed a tool to automatically segment these volumetric images, using manually segmented samples to tune and train a machine learning model. For a set of four specimens of ancient Egyptian animal mummies we achieve an overall accuracy of 94–98% when compared with manually segmented slices, approaching the results of off-the-shelf commercial software using deep learning (97–99%) at much lower complexity. A qualitative analysis of the segmented output shows that our results are close in terms of usability to those from deep learning, justifying the use of these techniques.

show abstract

“…For example, Hati et al [5] segmented an entire human mummy into jewels, body, bandages, and the wooden support frame, using a combination of voxel intensity and the manual selection of geodesic shapes. Low resolution volumetric scans of mummies were considered in [6,7], with segmentation performed one slice at a time using classical machine learning techniques. In both works, ridged artifacts and loss of detail were evident in the results obtained.…”

Section: Introductionmentioning

confidence: 99%

Automated segmentation of microtomography imaging of Egyptian mummies

Tanti,

Berruyer,

Tafforeau

et al. 2021

Preprint

View full text Add to dashboard Cite

Propagation Phase Contrast Synchrotron Microtomography (PPC-SRµCT) is the gold standard for non-invasive and non-destructive access to internal structures of archaeological remains. In this analysis, the virtual specimen needs to be segmented to separate different parts or materials, a process that normally requires considerable human effort. In the Automated SEgmentation of Microtomography Imaging (ASEMI) project, we developed a tool to automatically segment these volumetric images, using manually segmented samples to tune and train a machine learning model. For a set of four specimens of ancient Egyptian animal mummies we achieve an overall accuracy of 94-98% when compared with manually segmented slices, approaching the results of off-the-shelf commercial software using deep learning (97-99%) at much lower complexity. A qualitative analysis of the segmented output shows that our results are close in term of usability to those from deep learning, justifying the use of these techniques.

show abstract

A machine learning based approach to the segmentation of micro CT data in archaeological and evolutionary sciences

Cited by 3 publications

References 77 publications

A Shrewd Inspection of Vertebral Regionalization in Large Shrews (Soricidae: Crocidurinae)

A Shrewd Inspection of Vertebral Regionalization in Large Shrews (Soricidae: Crocidurinae)

Automated segmentation of microtomography imaging of Egyptian mummies

Automated segmentation of microtomography imaging of Egyptian mummies

Contact Info

Product

Resources

About