Iterative Next Boundary Detection for Instance Segmentation of Tree Rings in Microscopy Images of Shrub Cross Sections

TL;DR

This paper introduces INBD, an iterative method for precise instance segmentation of tree rings in microscopy images, effectively modeling growth direction and outperforming generic segmentation methods.

Contribution

The paper presents a novel iterative boundary detection approach specifically designed for tree ring segmentation, incorporating natural growth modeling and chronological order awareness.

Findings

INBD outperforms generic segmentation methods on the dataset.

It effectively models the growth direction of tree rings.

The method incorporates a built-in notion of chronological order.

Abstract

We address the problem of detecting tree rings in microscopy images of shrub cross sections. This can be regarded as a special case of the instance segmentation task with several unique challenges such as the concentric circular ring shape of the objects and high precision requirements that result in inadequate performance of existing methods. We propose a new iterative method which we term Iterative Next Boundary Detection (INBD). It intuitively models the natural growth direction, starting from the center of the shrub cross section and detecting the next ring boundary in each iteration step. In our experiments, INBD shows superior performance to generic instance segmentation methods and is the only one with a built-in notion of chronological order. Our dataset and source code are available at http://github.com/alexander-g/INBD.