In this work, we propose "Residual Attention Network", a convolutional neural
network using attention mechanism which can incorporate with state-of-art feed
forward network architecture in an end-to-end training fashion. Our Residual
Attention Network is built by stacking Attention Modules which generate
attention-aware features. The attention-aware features from different modules
change adaptively as layers going deeper. Inside each Attention Module,
bottom-up top-down feedforward structure is used to unfold the feedforward and
feedback attention process into a single feedforward process. Importantly, we
propose attention residual learning to train very deep Residual Attention
Networks which can be easily scaled up to hundreds of layers. Extensive
analyses are conducted on CIFAR-10 and CIFAR-100 datasets to verify the
effectiveness of every module mentioned above. Our Residual Attention Network
achieves state-of-the-art object recognition performance on three benchmark
datasets including CIFAR-10 (3.90% error), CIFAR-100 (20.45% error) and
ImageNet (4.8% single model and single crop, top-5 error). Note that, our
method achieves 0.6% top-1 accuracy improvement with 46% trunk depth and 69%
forward FLOPs comparing to ResNet-200. The experiment also demonstrates that
our network is robust against noisy labels.
Описание
Residual Attention Network for Image Classification
%0 Generic
%1 wang2017residual
%A Wang, Fei
%A Jiang, Mengqing
%A Qian, Chen
%A Yang, Shuo
%A Li, Cheng
%A Zhang, Honggang
%A Wang, Xiaogang
%A Tang, Xiaoou
%D 2017
%K cs.CV
%T Residual Attention Network for Image Classification
%U http://arxiv.org/abs/1704.06904
%X In this work, we propose "Residual Attention Network", a convolutional neural
network using attention mechanism which can incorporate with state-of-art feed
forward network architecture in an end-to-end training fashion. Our Residual
Attention Network is built by stacking Attention Modules which generate
attention-aware features. The attention-aware features from different modules
change adaptively as layers going deeper. Inside each Attention Module,
bottom-up top-down feedforward structure is used to unfold the feedforward and
feedback attention process into a single feedforward process. Importantly, we
propose attention residual learning to train very deep Residual Attention
Networks which can be easily scaled up to hundreds of layers. Extensive
analyses are conducted on CIFAR-10 and CIFAR-100 datasets to verify the
effectiveness of every module mentioned above. Our Residual Attention Network
achieves state-of-the-art object recognition performance on three benchmark
datasets including CIFAR-10 (3.90% error), CIFAR-100 (20.45% error) and
ImageNet (4.8% single model and single crop, top-5 error). Note that, our
method achieves 0.6% top-1 accuracy improvement with 46% trunk depth and 69%
forward FLOPs comparing to ResNet-200. The experiment also demonstrates that
our network is robust against noisy labels.
@misc{wang2017residual,
abstract = {In this work, we propose "Residual Attention Network", a convolutional neural
network using attention mechanism which can incorporate with state-of-art feed
forward network architecture in an end-to-end training fashion. Our Residual
Attention Network is built by stacking Attention Modules which generate
attention-aware features. The attention-aware features from different modules
change adaptively as layers going deeper. Inside each Attention Module,
bottom-up top-down feedforward structure is used to unfold the feedforward and
feedback attention process into a single feedforward process. Importantly, we
propose attention residual learning to train very deep Residual Attention
Networks which can be easily scaled up to hundreds of layers. Extensive
analyses are conducted on CIFAR-10 and CIFAR-100 datasets to verify the
effectiveness of every module mentioned above. Our Residual Attention Network
achieves state-of-the-art object recognition performance on three benchmark
datasets including CIFAR-10 (3.90% error), CIFAR-100 (20.45% error) and
ImageNet (4.8% single model and single crop, top-5 error). Note that, our
method achieves 0.6% top-1 accuracy improvement with 46% trunk depth and 69%
forward FLOPs comparing to ResNet-200. The experiment also demonstrates that
our network is robust against noisy labels.},
added-at = {2022-04-06T11:27:51.000+0200},
author = {Wang, Fei and Jiang, Mengqing and Qian, Chen and Yang, Shuo and Li, Cheng and Zhang, Honggang and Wang, Xiaogang and Tang, Xiaoou},
biburl = {https://www.bibsonomy.org/bibtex/2714c765ef6b60d0bf1449988f76c3024/aerover},
description = {Residual Attention Network for Image Classification},
interhash = {45bd7d65bac9a908b47e5beff49a7abf},
intrahash = {714c765ef6b60d0bf1449988f76c3024},
keywords = {cs.CV},
note = {cite arxiv:1704.06904Comment: accepted to CVPR2017},
timestamp = {2022-04-06T11:27:51.000+0200},
title = {Residual Attention Network for Image Classification},
url = {http://arxiv.org/abs/1704.06904},
year = 2017
}