Mask Region-Based Convolutional Neural Network (R-CNN) Based Image Segmentation of Rays in Softwoods

The current study aimed to verify the image segmentation ability of rays in tangential thin sections conifers using artificial intelligence technology. applied model was Mask region-based convolutional neural network (Mask R-CNN) and softwoods (viz. Picea jezoensis, Larix gmelinii, Abies nephrolepis, koreana, Ginkgo biloba, Taxus cuspidata, Cryptomeria japonica, Cedrus deodara, Pinus koraiensis) were selected for study. To take digital pictures, thickness 10–15 μm cut a microtome, then stained 1:1 mixture 0.5% astra blue 1% safranin. In images, as detection objects, Computer Vision Annotation Tool used annotate training images taken from woods. performance R-CNN select high 0.837 mean average precision saving time more than half that required Ground Truth. During analysis process, however, division into two or occurred. This caused some errors measurement ray height. improve processing algorithms, further work on combining fragments one segment, increasing boundary between neighboring tissues is required.