Fully Convolutional Networks for Semantic Segmentation

TL;DR

This paper introduces fully convolutional networks for semantic segmentation, enabling end-to-end pixel-wise predictions that significantly improve accuracy and inference speed on multiple datasets.

Contribution

It adapts classification networks into fully convolutional architectures and proposes a skip architecture for detailed segmentation, advancing the state-of-the-art.

Findings

30% relative improvement on PASCAL VOC 2012

Achieves 67.2% mean IU on PASCAL VOC 2012

Inference time of one-tenth of a second per image

Abstract

Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, improve on the previous best result in semantic segmentation. Our key insight is to build "fully convolutional" networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning. We define and detail the space of fully convolutional networks, explain their application to spatially dense prediction tasks, and draw connections to prior models. We adapt contemporary classification networks (AlexNet, the VGG net, and GoogLeNet) into fully convolutional networks and transfer their learned representations by fine-tuning to the segmentation task. We then define a skip architecture that combines semantic information from a deep, coarse layer with appearance information from a shallow, fine layer to produce accurate and detailed segmentations. Our fully convolutional network achieves improved segmentation of PASCAL VOC (30% relative improvement to 67.2% mean IU on 2012), NYUDv2, SIFT Flow, and PASCAL-Context, while inference takes one tenth of a second for a typical image.