Convolution Neural Networks for Semantic Segmentation: Application to Small Datasets of Biomedical Images

TL;DR

This thesis investigates how convolutional neural networks perform in biomedical image segmentation on small datasets, exploring various architectures and parameters to identify optimal configurations and underlying regularities.

Contribution

It systematically compares CNN architectures and hyper-parameters for small biomedical datasets, revealing preferred configurations and regularities in segmentation performance.

Findings

Certain network configurations outperform others on small datasets

Optimal hyper-parameters improve segmentation accuracy

Regularities in network performance identified

Abstract

This thesis studies how the segmentation results, produced by convolutional neural networks (CNN), is different from each other when applied to small biomedical datasets. We use different architectures, parameters and hyper-parameters, trying to find out the better configurations for our task, and trying to find out underlying regularities. Two working datasets are from biomedical area of research. We conducted a lot of experiments with the two types of networks and the received results have shown the preference of some conditions of experiments and parameters of the networks over the others. All testing results are given in the tables and some selected resulting graphs and segmentation predictions are shown for better illustration.