A key feature of ROXAS is to process images of entire samples in one step, including up to 1,000,000 cells and >100 annual rings. During automatic image analysis, ROXAS corrects image and samples deficiencies including poor contrast and ripped-off cell walls extending into the cell lumina. Detected anatomical features are represented and saved as vector overlays. Efficient cell filtering and editing tools allow to improve results from automatic ROXAS analysis. The automatic analysis can be executed in batch mode to limit user waiting time to a minimum (full feature overview and screenshots).
ROXAS produces and saves >120 output parameters on the level of the entire image, each annual ring and each individual cell, including annual ring width, cell lumen area and diameter, cell wall thickness (tangential and radial), cell position in the ring, vessel grouping, hydraulic conductance, Mork’s index and the ‘bending resistance’ (t/b)2. ROXAS can also measure characteristics of conifer needles including needle surface and length (detailed parameter list). Data output is saved into MS Excel and tab-delimited text files. An existing analysis can be re-loaded at any time for quality control and further editing.
More than 150 customizable settings used during automatic analysis are saved in configuration files. This allows to tailor the automatic analysis and data output to the specific project (e.g., anatomy of the analyzed species, quality and staining of the anatomical samples, image quality). A large collection of pre-defined configurations for distinct samples is available in the ROXAS Configuration Library.
ROXAS has a couple of requirements and dependencies. It supports common image formats such as JPG, JPG2000 and TIFF. It requires a Windows operating system (on PC or MAC). And it depends on the commercial software Image-Pro Plus v6.1-7 (Media Cybernetics) for the general image operations. There is currently no alternative to the latter dependency (full list of requirements).
Note that the quality of the produced data depends much on the quality of the analyzed images (besides proper configurations and sufficient manual editing). With a poor-quality anatomical sample and image the manual editing may easily require 10x longer and more, while the achieved data quality may still be insufficient. Some advice and further reading on the proper sample preparation and digitizing can be obtained in the 'Sample Preparation and Image Acquisition' sub-section.
N.B.: The name ‘ROXAS’ goes back to the prototype version and refers to ‘Root xylem analysis system’ – in those days ROXAS was designed to only analyze images of herbaceous root cross-sections.