FibeR-CNN: Expanding Mask R-CNN to improve image-based fiber analysis

Fiber-shaped materials (e.g. carbon nano tubes) are of great relevance, due to their unique properties but also the health risk they can impose. Unfortunately, image-based analysis fibers still involves manual annotation, which is a time-consuming and costly process. We therefore propose use region-based convolutional neural networks (R-CNNs) automate this task. Mask R-CNN, most widely used R-CNN for semantic segmentation tasks, prone errors when it comes fiber-shaped objects. Hence, new architecture - FibeR-CNN introduced validated. combines two established architectures (Mask Keypoint R-CNN) adds additional network heads prediction fiber widths lengths. As result, able surpass mean average precision by 33 % (11 percentage points) on novel test data set images.