# Efficient Ladder-style DenseNets for Semantic Segmentation of Large Images

## TL;DR

This paper introduces an efficient DenseNet-based ladder architecture for semantic segmentation that reduces memory usage, enabling training on high-resolution images with improved accuracy and speed.

## Contribution

A novel ladder-style DenseNet architecture that minimizes feature map caching, allowing high-resolution training with fewer parameters and better performance.

## Key findings

- Outperforms state-of-the-art in accuracy and speed
- Enables training on megapixel images with commodity hardware
- Uses fewer parameters than competing models

## Abstract

Recent progress of deep image classification models has provided great potential to improve state-of-the-art performance in related computer vision tasks. However, the transition to semantic segmentation is hampered by strict memory limitations of contemporary GPUs. The extent of feature map caching required by convolutional backprop poses significant challenges even for moderately sized Pascal images, while requiring careful architectural considerations when the source resolution is in the megapixel range. To address these concerns, we propose a novel DenseNet-based ladder-style architecture which features high modelling power and a very lean upsampling datapath. We also propose to substantially reduce the extent of feature map caching by exploiting inherent spatial efficiency of the DenseNet feature extractor. The resulting models deliver high performance with fewer parameters than competitive approaches, and allow training at megapixel resolution on commodity hardware. The presented experimental results outperform the state-of-the-art in terms of prediction accuracy and execution speed on Cityscapes, Pascal VOC 2012, CamVid and ROB 2018 datasets. Source code will be released upon publication.