Fully convolutional networks for semantic segmentation

TLDR

Convolutional networks are powerful visual models that produce hierarchical feature representations. The authors aim to develop fully convolutional networks that accept arbitrary‑sized inputs and generate correspondingly sized outputs for efficient dense prediction. They convert existing classification architectures such as AlexNet, VGG, and GoogLeNet into fully convolutional forms, fine‑tune them for segmentation, and introduce a skip architecture that fuses deep semantic and shallow appearance cues. The resulting fully convolutional network surpasses prior state‑of‑the‑art on PASCAL VOC, NYUDv2, and SIFT Flow—achieving a 62.2% mean IU on 2012 VOC with a 20% relative improvement—while performing inference in under one‑fifth 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, exceed the state-of-the-art 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 [20], the VGG net [31], and GoogLeNet [32]) into fully convolutional networks and transfer their learned representations by fine-tuning [3] 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 state-of-the-art segmentation of PASCAL VOC (20% relative improvement to 62.2% mean IU on 2012), NYUDv2, and SIFT Flow, while inference takes less than one fifth of a second for a typical image.