Comparing Normalization Methods for Limited Batch Size Segmentation Neural Networks

TL;DR

This paper evaluates normalization methods for 3D CT spine segmentation with limited batch sizes, finding Instance Normalization to be most effective and computationally efficient.

Contribution

It demonstrates that Instance Normalization outperforms other methods in limited batch size segmentation tasks, achieving high accuracy with efficient computation.

Findings

Instance Normalization achieved Dice coefficient = 0.96

Outperforms other normalization methods in limited batch scenarios

Computationally efficient compared to no normalization

Abstract

The widespread use of Batch Normalization has enabled training deeper neural networks with more stable and faster results. However, the Batch Normalization works best using large batch size during training and as the state-of-the-art segmentation convolutional neural network architectures are very memory demanding, large batch size is often impossible to achieve on current hardware. We evaluate the alternative normalization methods proposed to solve this issue on a problem of binary spine segmentation from 3D CT scan. Our results show the effectiveness of Instance Normalization in the limited batch size neural network training environment. Out of all the compared methods the Instance Normalization achieved the highest result with Dice coefficient = 0.96 which is comparable to our previous results achieved by deeper network with longer training time. We also show that the Instance Normalization implementation used in this experiment is computational time efficient when compared to the network without any normalization method.