Rethinking the Inception Architecture for Computer Vision
Christian Szegedy
Google Inc.
szegedy@google.com
Vincent Vanhoucke
vanhoucke@google.com
Sergey Ioffe
sioffe@google.com
Jonathon Shlens
shlens@google.com
Zbigniew Wojna
University College London
zbigniewwojna@gmail.com
Abstract
Convolutional networks are at the core of most state-
of-the-art computer vision solutions for a wide variety of
tasks. Since 2014 very deep convolutional networks started
to become mainstream, yielding substantial gains in vari-
ous benchmarks. Although increased model size and com-
putational cost tend to translate to immediate quality gains
for most tasks (as long as enough labeled data is provided
for training), computational efficiency and low parameter
count are still enabling factors for various use cases such as
mobile vision and big-data scenarios. Here we are explor-
ing ways to scale up networks in ways that aim at utilizing
the added computation as efficiently as possible by suitably
factorized convolutions and aggressive regularization. We
benchmark our methods on the ILSVRC 2012 classification
challenge validation set demonstrate substantial gains over
the state of the art: 21.2% top-1 and 5.6% top-5 error for
single frame evaluation using a network with a computa-
tional cost of 5 billion multiply-adds per inference and with
using less than 25 million parameters. With an ensemble of
4 models and multi-crop evaluation, we report 3.5% top-5
error and 17.3% top-1 error.
1. Introduction
Since the 2012 ImageNet competition [16] winning en-
try by Krizhevsky et al [9], their network “AlexNet” has
been successfully applied to a larger variety of computer
vision tasks, for example to object-detection [5], segmen-
tation [12], human pose estimation [22], video classifica-
tion [8], object tracking [23], and superresolution [3].
These successes spurred a new line of research that fo-
cused on finding higher performing convolutional neural
networks. Starting in 2014, the quality of network architec-
tures significantly improved by utilizing deeper and wider
networks. VGGNet [18] and GoogLeNet [20] yielded simi-
larly high performance in the 2014 ILSVRC [16] classifica-
tion challenge. One interesting observation was that gains
in the classification performance tend to transfer to signifi-
cant quality gains in a wide variety of application domains.
This means that architectural improvements in deep con-
volutional architecture can be utilized for improving perfor-
mance for most other computer vision tasks that are increas-
ingly reliant on high quality, learned visual features. Also,
improvements in the network quality resulted in new appli-
cation domains for convolutional networks in cases where
AlexNet features could not compete with hand engineered,
crafted solutions, e.g. proposal generation in detection[4].
Although VGGNet [18] has the compelling feature of
architectural simplicity, this comes at a high cost: evalu-
ating the network requires a lot of computation. On the
other hand, the Inception architecture of GoogLeNet [20]
was also designed to perform well even under strict con-
straints on memory and computational budget. For exam-
ple, GoogleNet employed only 5 million parameters, which
represented a 12⇥ reduction with respect to its predeces-
sor AlexNet, which used 60 million parameters. Further-
more, VGGNet employed about 3x more parameters than
AlexNet.
The computational cost of Inception is also much lower
than VGGNet or its higher performing successors [6]. This
has made it feasible to utilize Inception networks in big-data
scenarios[17], [13], where huge amount of data needed to
be processed at reasonable cost or scenarios where memory
or computational capacity is inherently limited, for example
in mobile vision settings. It is certainly possible to mitigate
parts of these issues by applying specialized solutions to tar-
get memory use [2], [15] or by optimizing the execution of
certain operations via computational tricks [10]. However,
these methods add extra complexity. Furthermore, these
methods could be applied to optimize the Inception archi-
tecture as well, widening the efficiency gap again.
Still, the complexity of the Inception architecture makes
1
arXiv:1512.00567v3 [cs.CV] 11 Dec 2015
资源评论
不脱发的程序猿
- 粉丝: 26w+
- 资源: 5889
最新资源
- MATLAB代码：基于储能电站服务的冷热电多微网系统双层优化配置 关键词：储能电站 共享储能电站 冷热电多微网 双层优化配置 参考文档：《基于储能电站服务的冷热电多微网系统双层优化配置》完全复
- CarSim or TruckSim与Simulink联合仿真，使用键盘控制车辆加速，减速，转向，包含前进与后 档位切，支持自定义按键功能，支持拓展 提供carsim参数配置文件，导入即可运行 提供s
- COMSOL:AlGaAs纳米天线倍频产生（二次谐波非线性效应SHG）三维模型
- 内容介绍：考虑风光不确定性的微电网优化，机会约束描述风光不确定性，采用概率序列处理不确定问题 代码来源：原作者，非倒卖 代码说明：该代码主要贡献为不确定性处理和优化，可供大家学习参考
- 电动汽车 双层优化 选址定容 输配协同 时空优化 MATLAB代码：基于双层优化的电动汽车优化调度研究 参考文档：考虑大规模电动汽车接入电网的双层优化调度策略 仿真平台：MATLAB+CPLEX
- 基于拉丁超立方抽样的风，光，负荷场景生成方法 风电功率场景生成 ，光伏功率场景生成，负荷场景生成 通过后向场景削减BR得到典型场景及其概率 提供参考文献，完美复现 语言：MATLAB
- 西门子博途热力发电厂汽轮机控制系统程序画面案例，采用西门子S7-1516cpu,画面采用经典WINCC7.5 文件内容包含工艺流程图，电仪设计EPLAN图，及控制逻辑等 博图版本V16
- 51单片机PID法设计的开关电源Proteus仿真 功能描述如下： 1、220交流电输入整流滤波得到18 V直流，通过7805得到5v直流电给单片机供电； 2、18 V直流电经过开关变电路(buck
- 通过node Red的OPC UA节点，使用OPC UA server节点建立服务，量NX MCD作为客户端实现数据的交互 node Red+NX MCD轻松实现可视化界面的开发，实现远程控制，数据
- IM即时通讯APP源码搭建语音视频聊天交友软件
- Carsim和simulink联合仿真车辆状态估计 估计的状态为:横摆角速度，质心侧偏角，纵向车速，侧向车速 先基于滑模观测器SMO估计轮胎的纵向力和侧向力 ，再基于无迹卡尔曼UKF和容积卡尔曼CKF
- 三菱Q系列程序 三菱Q06UDV程序，搭配三菱MR-JE伺服，QD77MS16运动控制模块，QD70P8运动模块控制24轴运动，QX42.QY42P等输入输出模块 全自动iPad撕胶贴二维
- 负荷需求响应，matlab 在基于价格的需求侧管理模型研究中，首要任务便是建立负荷对价格的响应模型 有的文献中建立了价格型需求响应功率对电价的响应模型，认为两者之间是简单的线性关系 也有文献忽略了
- BYVIN 比德文知名厂家电动四轮车控制器代码，包含PCB文件，pdf原理图，代码齐全，风格很好
- 质子交膜燃料电池pemfc热管理水管理进气控制 固体氧化物燃料电池sofc 模糊控制，pid控制，模糊pid控制 自抗扰控制 bp-pid rbf-pid ga-pid pso-pid matlab
- 利用最小支持向量机LSSVM做拟合预测建模，数据要求是多列输入单列输出做拟合预测建模，程序内注释详细，直接替数据就可以用
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
