Boosted Training of Convolutional Neural Networks for Multi-Class Segmentation

TL;DR

This paper introduces a boosted training approach for convolutional neural networks that enhances segmentation accuracy and training efficiency on large volumetric datasets by focusing on difficult regions and adaptively tuning learning rates.

Contribution

It presents a novel boosted sampling scheme using error maps, an adaptive learning rate schedule, and achieves state-of-the-art results on a benchmark dataset.

Findings

Faster training with improved segmentation performance

Effective focus on difficult regions during training

Achieved new state-of-the-art results on VISCERAL benchmark

Abstract

Training deep neural networks on large and sparse datasets is still challenging and can require large amounts of computation and memory. In this work, we address the task of performing semantic segmentation on large volumetric data sets, such as CT scans. Our contribution is threefold: 1) We propose a boosted sampling scheme that uses a-posterior error maps, generated throughout training, to focus sampling on difficult regions, resulting in a more informative loss. This results in a significant training speed up and improves learning performance for image segmentation. 2) We propose a novel algorithm for boosting the SGD learning rate schedule by adaptively increasing and lowering the learning rate, avoiding the need for extensive hyperparameter tuning. 3) We show that our method is able to attain new state-of-the-art results on the VISCERAL Anatomy benchmark.