作者：Francois Chollet（深度学习框架Keras作者）

作者单位：Google

发布时间：2017

发布期刊/会议：CVPR

论文全称：Xception: Deep Learning With Depthwise Separable Convolutions

论文地址：

CVPR 2017 Open Access Repository

论文代码：

keras/xception.py at master · keras-team/keras

地位：整个Inception系列的封神之作，是Inception系列的极致版本，使用 Depthwise Separable Convlution 将空间信息与通道信息彻底解耦

个人理解

• **创新点：**将深度可分离卷积和Inception模块相结合，设计出Xception网络，并提出了可能的研究方向。
• **为什么：**Inception系列中的Inception模块还可以继续改进，以设计出最优的模型
• **怎么做：**简而言之，Inception模块就是在常规卷积（一个kernel处理所有channel）和深度可分离卷积（一个kernel处理一个channel）之间的模块，即一个kernl处理n个channel，即作者通过调整每个kernel处理的channel数或比例来设计不同的Inception模块，而因为时间关系，作者直接取了最极致（extreme）的一种方式，即一个kernel处理一个channel。

一、摘要

We present an interpretation of Inception modules in convolutional neural networks as being an intermediate step in-between regular convolution and the depthwise separable convolution operation (a depthwise convolution followed by a pointwise convolution). In this light, a depthwise separable convolution can be understood as an Inception module with a maximally large number of towers. This observation leads us to propose a novel deep convolutional neural network architecture inspired by Inception, where Inception modules have been replaced with depthwise separable convolutions. We show that this architecture, dubbed Xception, slightly outperforms Inception V3 on the ImageNet dataset (which Inception V3 was designed for), and signifificantly outperforms Inception V3 on a larger image classifification dataset comprising 350 million images and 17,000 classes. Since the Xception architecture has the same number of parameters as Inception V3, the performance gains are not due to increased capacity but rather to a more effificient use of model parameters.

• Inception 结构是介于传统卷积和深度可分离卷积的中间形态，Xception 彻底解耦为深度可分离卷积。