Convolutional Neural Networks(3):Convolution and Channels
在CNN(1)和CNN(2)两篇文章中,主要说明的是CNN的基本架构和权值共享(Weight Sharing),本文则重点介绍卷积的部分。
首先,在卷积之前,我们的数据是4D的tensor(width,height,channels,batch),在CNN(1):Architecture一文中,曾经提到过。这里的通道,和之前所说的depth是一个概念,例如一张Grey Scale Image,其通道数为1;RGB图的通道数为3。
而事实上,kernel也是有channel的,并且其数量与输入tensor的通道数一致。如下图的RGB图像,有3个通道,在卷积过程中我们使用5个kernel。
那么,到底有多少feature map?有很长一段时间,我的答案是15个feature map。但实际上,答案是5个(和kernel的数量一致)。如图示,Kernel也有3个通道,分别与RGB三个通道进行卷积后,生成了3个单通道卷积结果,而后三个结果相加得到feature map。而在此卷积层中,kernel是我们唯独需要“学习”的参数,也就是3*3*3*5=45个参数。实际上,通过训练kernel会学习到应该过滤图片中的哪些特征。
Convolutional Neural Networks(3):Convolution and Channels的更多相关文章
- [CS231n-CNN] Convolutional Neural Networks: architectures, convolution / pooling layers
课程主页:http://cs231n.stanford.edu/ 参考: 细说卷积神经网络:http://blog.csdn.net/han_xiaoyang/article/details/ ...
- [转] Understanding Convolutional Neural Networks for NLP
http://www.wildml.com/2015/11/understanding-convolutional-neural-networks-for-nlp/ 讲CNN以及其在NLP的应用,非常 ...
- 课程四(Convolutional Neural Networks),第一周(Foundations of Convolutional Neural Networks) —— 3.Programming assignments:Convolutional Model: application
Convolutional Neural Networks: Application Welcome to Course 4's second assignment! In this notebook ...
- 课程四(Convolutional Neural Networks),第一周(Foundations of Convolutional Neural Networks) —— 2.Programming assignments:Convolutional Model: step by step
Convolutional Neural Networks: Step by Step Welcome to Course 4's first assignment! In this assignme ...
- Convolutional Neural Networks: Step by Step
Andrew Ng deeplearning courese-4:Convolutional Neural Network Convolutional Neural Networks: Step by ...
- Convolutional Neural Networks: Application
Andrew Ng deeplearning courese-4:Convolutional Neural Network Convolutional Neural Networks: Step by ...
- (zhuan) Building Convolutional Neural Networks with Tensorflow
Ahmet Taspinar Home About Contact Building Convolutional Neural Networks with Tensorflow Posted on a ...
- [转]An Intuitive Explanation of Convolutional Neural Networks
An Intuitive Explanation of Convolutional Neural Networks https://ujjwalkarn.me/2016/08/11/intuitive ...
- Understanding Convolutional Neural Networks for NLP
When we hear about Convolutional Neural Network (CNNs), we typically think of Computer Vision. CNNs ...
随机推荐
- 2019 Multi-University Training Contest 2 - 1008 - Harmonious Army - 最大流
http://acm.hdu.edu.cn/showproblem.php?pid=6598 一开始就觉得是网络流,但是一直都不会怎么建图. 这里要考虑. 每一组边(u,v,a,b,c)建立如下的连接 ...
- Leetcode Lect7 哈希表
传统的哈希表 对于长度为n的哈希表,它的存储过程如下: 根据 key 计算出它的哈希值 h=hash(key) 假设箱子的个数为 n,那么这个键值对应该放在第 (h % n) 个箱子中 如果该箱子中已 ...
- maven 打包Scala代码到jar包
idea的pom.xml文件配置 <dependencies> <dependency> <groupId>org.scala-lang</groupId&g ...
- C语言如何使输出的数字对齐
右对齐%numd(num是位数,比如按5位数的长度输出,num为正数则右对齐) #include <stdio.h> int main() { printf(, ); printf(, ) ...
- 线程局部变量ThreadLocal实现原理
ThreadLocal,即线程局部变量,用来为每一个使用它的线程维护一个独立的变量副本.这种变量只在线程的生命周期内有效.并且与锁机制那种以时间换取空间的做法不同,ThreadLocal没有任何锁机制 ...
- Apache Mesos1.0.1 编译安装部署教程(ubuntu)
参考资料 官方文档:http://mesos.apache.org/documentation 中文翻译:http://mesos.mydoc.io/ GitHub:https://github.co ...
- Tomcat部署虚拟主机
使用Tomcat部署加密虚拟主机,实现: a.使用www.a.com域名访问的页面根路径为/usr/local/tomcat/a/base b.使用www.b.com域名访问的页面根路径为/usr/l ...
- nginx中break和last的区别
一.环境准备 资源文件创建 mkdir -p /opt/tmp/wqy/test/aaa mkdir -p /opt/tmp/wqy/test/bbb echo "aaa" > ...
- no hash tools
import itertools class Set(list): def __init__(self, params): super(Set, self).__init__() ...
- 【LGR-062】洛谷10月月赛 III div.2 (A-C)
前言 100+100+46+0=246pts 300多名 以后每次比赛都要有进步哦!qwq 小D与笔试 水题 Code #include<algorithm> #include<io ...