Journal: Philosophical Transactions of the Royal Society B: Biological Sciences

Article Title: Towards a morphological metric of assemblage dynamics in the fossil record: a test case using planktonic foraminifera

doi: 10.1098/rstb.2015.0227

Figure Lengend Snippet: Visual pipeline illustrating the steps involved in 3D mesh extraction using the run3dmorph software. ( a ) Z -stacks of each individual object (taken at varying focal planes of known height above the object) are processed using the Stack Focuser plug-in for ImageJ/FIJI , resulting in a focused image of the object and a height map (built using an 11 × 11 pixel kernel size). ( b ) The focused image and height map are rescaled such that each pixel has a height and width of 1 µm, and the 2D outline of the focused image is extracted using the run2dmorph software (see https://github.com/Hull-Lab ). Each pixel of the binary 2D outline image is then multiplied against the corresponding pixel in the height map (element-wise multiplication). This effectively deletes background noise and results in a cleaned-up height map ( c ). The greyscale value of each pixel in the height map is then used, in conjunction with the distance between each z -stack slice, to back-calculate the real-world height of each pixel and generate an unfiltered 3D mesh ( d ). High and low outlier noise is then filtered from the 3D mesh using a custom neighbourhood pixel-averaging algorithm (using a user-defined n × n pixel kernel, where n is a positive odd integer). Vertex and face coordinates are then extracted from the cleaned 3D mesh and outputted in standard 3D ASCII formats (OBJ and OFF).

Article Snippet: The first two components of the pipeline ( segment and focus ) are written in Python, a free programming language that runs across platforms. run2dmorph and run3dmorph currently execute in MATLAB (version 2015b or above), a proprietary software, but will be ported into Python in future versions.

Techniques: Extraction, Software

Journal: Philosophical Transactions of the Royal Society B: Biological Sciences

Article Title: Towards a morphological metric of assemblage dynamics in the fossil record: a test case using planktonic foraminifera

doi: 10.1098/rstb.2015.0227

Figure Lengend Snippet: Illustration of idealized base shapes, used by run3dmorph for estimating surface area and volume of the complete object. ( a ) Irregular cone base; ( b ) irregular cylinder base; ( c ) spheroidal dome base.

Article Snippet: The first two components of the pipeline ( segment and focus ) are written in Python, a free programming language that runs across platforms. run2dmorph and run3dmorph currently execute in MATLAB (version 2015b or above), a proprietary software, but will be ported into Python in future versions.

Techniques:

Journal: Philosophical Transactions of the Royal Society B: Biological Sciences

Article Title: Towards a morphological metric of assemblage dynamics in the fossil record: a test case using planktonic foraminifera

doi: 10.1098/rstb.2015.0227

Figure Lengend Snippet: Visual comparison of data types: Tohoku University 3D specimens from CT and semi-3D half-hulls extracted using the run3dmorph software on stacked microscopic images. Three examples specimens are shown: Trilobatus sacculifer , Truncorotalia truncatulinoides and Neogloboquadrina dutertrei . The Tohoku University specimens were digitized using X-ray CT at 5 µm resolution. Run3dmorph -extracted 3D-meshes are shown next to their corresponding focused 2D-image.

Article Snippet: The first two components of the pipeline ( segment and focus ) are written in Python, a free programming language that runs across platforms. run2dmorph and run3dmorph currently execute in MATLAB (version 2015b or above), a proprietary software, but will be ported into Python in future versions.

Techniques: Comparison, Software

Journal: Philosophical Transactions of the Royal Society B: Biological Sciences

Article Title: Towards a morphological metric of assemblage dynamics in the fossil record: a test case using planktonic foraminifera

doi: 10.1098/rstb.2015.0227

Figure Lengend Snippet: Exploration of the reproducibility tests of the Orbulina universa individual ( a ) The PC1 versus PC2 coordinates for O. universa from figure 8 c . The points highlighted in red are outliers for which the run3dmorph -extracted semi-3D half-hulls exhibited pathologies (see text). ( b – d ) Extracted mesh, focused image and height map showcasing pathologies arising during z -stack focusing. In ( c ), the smeared, unfocused portions of the object are outlined in white. ( e – g ) The corresponding mesh, focused image and height map for a properly extracted O. orbulina individual.

Article Snippet: The first two components of the pipeline ( segment and focus ) are written in Python, a free programming language that runs across platforms. run2dmorph and run3dmorph currently execute in MATLAB (version 2015b or above), a proprietary software, but will be ported into Python in future versions.

Techniques: