Deep Residual Learning for Image Recognition

TL;DR

This paper introduces residual learning frameworks that enable training of substantially deeper neural networks, significantly improving accuracy and winning top prizes in major image recognition competitions.

Contribution

The paper proposes residual networks that reformulate layers as learning residual functions, making very deep networks easier to optimize and more accurate.

Findings

Residual networks outperform previous models on ImageNet with up to 152 layers.

Ensemble of residual nets achieves 3.57% error on ImageNet test set.

Deep residual nets improve object detection accuracy by 28% on COCO dataset.

Abstract

Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG nets but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.