Region Growing with Convolutional Neural Networks for Biomedical Image Segmentation

TL;DR

This paper introduces a CNN-based region growing method for biomedical image segmentation that iteratively expands predicted regions, achieving high accuracy with limited training data and preserving biological morphology.

Contribution

It presents a novel CNN-driven region growing approach that improves segmentation accuracy and morphological preservation over traditional fully convolutional methods.

Findings

Outperforms fully convolutional CNNs in accuracy on retinal blood vessel images.

Maintains biological morphology while using small training datasets.

Offers computational efficiency in segmentation tasks.

Abstract

In this paper we present a methodology that uses convolutional neural networks (CNNs) for segmentation by iteratively growing predicted mask regions in each coordinate direction. The CNN is used to predict class probability scores in a small neighborhood of the center pixel in a tile of an image. We use a threshold on the CNN probability scores to determine whether pixels are added to the region and the iteration continues until no new pixels are added to the region. Our method is able to achieve high segmentation accuracy and preserve biologically realistic morphological features while leveraging small amounts of training data and maintaining computational efficiency. Using retinal blood vessel images from the DRIVE database we found that our method is more accurate than a fully convolutional semantic segmentation CNN for several evaluation metrics.