X-Ray Tomography of Radiolarians
geometry of the radiolarian cage- like test. T e voxel size of the original data set was used to measure the lengths of the skeleton segments. Segment number and nodes (inter- secting points of segments) were also com puted. All these functions are available with the latest version of AMIRA (6.0). T e rendered models were con- verted to VRML fi les with AMIRA and then to STL fi les with nettfab. Several high-resolution models were 3D printed by Shapeways with a strong, fl exible polymer.
Results
Several radiolaria embedded in the nitrocellulose adhesive were scanned as a group ( Figure 1 ). A variety of forms were present, each perforated with diff erent patterns forming cage-like structures with polygonal outlines. Selected speci- mens were volume edited (cropped in three dimensions) and surface rendered as individual radiolarians ( Figures 2a–2d ). T ese rendered models exhibited sharply defi ned features, and their surface details were resolved well into the sub-µm range.
Figures 2a–2d : Selected radiolarian models exhibiting highly resolved, sharply defi ned surface elements. Models a, c, and d were volume-edited from the group exhibited in Figure 1 .
Figure 3 : Digital bisection of rendered models reveal medullary shells within the outer (cortical) shells. In some cases, a single medullar shell was held in place by struts oriented in three planes (3a). In other specimens, two medullary shells were present, each anchored to each other and to the cortical shell by multiple struts (3b).
Surface-rendered models were skeletonized. T is is an erosion or thinning of voxels so that only a string of voxels remains. T e resulting skeleton represents the essential
20
Digital bisection of rendered models revealed medullary shells held in place by struts connected to the outer cortical shell ( Figure 3a ). In some specimens, two medullary shells were present, one inside the other. The two inner shells were connected by struts, and both were connected to the outer shell by additional struts ( Figure 3b ). Several of these models were amenable to digital skeletonization [ 4 – 6 ]. This is a thinning algorithm that erodes voxels away from the modeled surface until only a single line of voxels remains (the medial axis). This worked well for specimens with cylindrical or near cylindrical segments, form- ing the polygons. In such cases the resulting skeleton accurately reflected the essential geometry of the radiolarian test ( Figure 4 ).
Topological features of the skeletons, such as nodes (intersec- tions of segments) and segment number, were computed. Segment length was calculated according to the voxel size
www.microscopy-today.com • 2015 September
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68