# 3DQ: Compact Quantized Neural Networks for Volumetric Whole Brain Segmentation

## TL;DR

The paper introduces 3DQ, a ternary quantization technique for 3D CNNs, achieving 16x compression while maintaining segmentation performance, thus reducing storage needs for brain segmentation models.

## Contribution

This work is the first to apply ternary quantization to 3D Fully Convolutional Neural Networks, enabling significant model compression without performance loss.

## Key findings

- Achieves 16x model compression with maintained accuracy.
- Effectively generalizes across 3D U-Net and V-Net architectures.
- Reduces storage requirements to a few MBytes for large models.

## Abstract

Model architectures have been dramatically increasing in size, improving performance at the cost of resource requirements. In this paper we propose 3DQ, a ternary quantization method, applied for the first time to 3D Fully Convolutional Neural Networks (F-CNNs), enabling 16x model compression while maintaining performance on par with full precision models. We extensively evaluate 3DQ on two datasets for the challenging task of whole brain segmentation. Additionally, we showcase our method's ability to generalize on two common 3D architectures, namely 3D U-Net and V-Net. Outperforming a variety of baselines, the proposed method is capable of compressing large 3D models to a few MBytes, alleviating the storage needs in space critical applications.