Bone geometry is commonly measured on computed tomographic (CT) and X-ray microtomographic (μCT) images. We obtained hundreds of CT, μCT and synchrotron μCT images of bones from diverse species that needed to be analysed remote from scanning hardware, but found that available software solutions were expensive, inflexible or methodologically opaque. We implemented standard bone measurements in a novel ImageJ plugin, BoneJ, with which we analysed trabecular bone, whole bones and osteocyte lacunae. BoneJ is open source and free for anyone to download, use, modify and distribute.

The local thickness described by Hildebrand and Rüesgsegger can characterize a 3D binary image of a complex structure such as a bone, a cell, or a paper fiber [1]. Such images are available from, for example, micro-computed tomography [2]. Let 3 R W be the structure. The local thickness at point W p is the diameter of the largest sphere that contains p and is completely inside the structure.The problem is to transform the 3D binary image into the 3D local thickness image.Several steps to are given in [1]. The first is a Euclidean distance transformation (EDT) The distance map at W q , ( ) q D map , is the distance from q to the nearest background point. The sphere of radius ( ) q D map centered at q is completely inside the structure. If p is within this sphere, then the local thickness at p, ( ) p τ , is at least ( ) q D map . Points q are scanned to find ( ) p τ by maximizing this bound. A further simplification is to first remove many of the points whose spheres of radius ( ) q D map are completely contained within the corresponding spheres of other points. Removing all of the redundant points would produce the distance ridge (DR) which is related to the skeleton and the medial axis [1, 4].A 3D local thickness procedure has been implemented in Java as a plugin for ImageJ [5]. The EDT uses the fast algorithm given by Saito and Toriwaki [6]. The DR computation is a simplified form of a template scheme [7]. Care was taken to optimize the use computational resources. No temporary 3D buffers are required. A related implementation in C was given previously by Coeurjolly [8]

No abstract

SummaryThis study presents the development of an ImageJ plugin for surface characterization. Based on gradient analysis, parameters, such as the gradient magnitude, orientation, mean resultant vector and surface area are derived. A comparative study of supercalendered (SC) papers was performed to verify the surface representations yielded by a laser profilometer. The surface representations of samples covered with carbon and gold were compared to untreated samples. The results confirm the suitability of gold coating for reducing the artefacts encountered on laser profilometry surface representations of paper. In addition, a complete scanning electron microscopy analysis is performed on the assessed samples to quantify the surface fraction covered by mineral fillers and to reveal the true 3D microstructure of SC surfaces. The influence of filler coverage and filler type on the gloss level of commercial SC papers is evaluated. The relationship between the surface topography, gloss and PPS roughness for a series of commercial SC papers is established.