Comparing Normalization Methods for Limited Batch Size Segmentation Neural Networks
This research provides a solution for medical image segmentation practitioners facing hardware limitations when training memory-intensive models, demonstrating that Instance Normalization can maintain performance with smaller batch sizes.
This paper addresses the challenge of training segmentation neural networks with limited batch sizes due to memory constraints. They compared various normalization methods for binary spine segmentation from 3D CT scans, finding that Instance Normalization achieved the highest Dice coefficient of 0.96, comparable to deeper networks trained with longer times.
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.