Dataset之Knifey-Spoony：Knifey-Spoony数据集的简介、下载、使用方法之详细攻略

Knifey-Spoony数据集的简介

##The Knifey-Spoony Data-Set

## Introduction

## Images

Knifey-Spoony数据集的下载

Knifey-Spoony数据集的使用方法

Knifey-Spoony数据集的简介

Knifey-Spoony数据集：含有刀子、叉子、汤匙，共计22M左右。Knifey-Spoony数据集由一个视频文件取出帧并转换成图像产生。训练集包含4170幅图像，测试集包含530幅图像。训练集中只有994张叉子的图像，却包含着1201张刀子的图像和1966张汤匙的图像。

##The Knifey-Spoony Data-Set

[Original repository on GitHub](https://github.com/Hvass-Labs/knifey-spoony)
Original author is [Magnus Erik Hvass Pedersen](http://www.hvass-labs.org)

## Introduction

* This is the Knifey-Spoony image data-set.
* The images show 3 types of objects: Cutlery knives, spoons and forks on a few different backgrounds.
* The classes are named: knifey, spoony and forky. (It's a spoof from The Simpsons on the 1980's movie Crocodile Dundee.)
* These images are used in [TensorFlow Tutorial #09](https://github.com/Hvass-Labs/TensorFlow-Tutorials/blob/master/08_Video_Data.ipynb) as an example of a classification problem.
* There is a [script](https://github.com/Hvass-Labs/TensorFlow-Tutorials/blob/master/convert.py) for converting videos to images so you can easily create your own data-set with thousands of images from just a few video-recordings.

*这是Knifey-Spoony图像数据集。
*图片显示了3种不同背景的物体：刀具、勺子和叉子。
*这些课程被命名为：刀叉，幽灵和叉子。（这是1980年代电影《The Simpsons》中辛普森一家的一个spoof。）
*这些图像在[tensorflow tutorial_09]中（https://github.com/hvass-labs/tensorflow-tutorials/blob/master/08_video_data.ipynb）用作分类问题的示例。
*有一个[脚本]（https://github.com/hvass-labs/tensorflow-tutorials/blob/master/convert.py）用于将视频转换为图像，这样您就可以轻松地创建自己的数据集，其中包含仅几段视频记录中的数千个图像。

## Images

* All images are 200 x 200 pixels with 3 colour channels.
* There is a total of 4700 jpg-images of which 530 are test-images with different backgrounds than the training-set.
* The knifey class has 1347 images total (137 images in the test-set).
* The spoony class has 2208 images total (242 images in the test-set).
* The forky class has 1145 images total (151 images in the test-set).

*所有图像均为200 x 200像素，带有3个彩色通道。
*总共有4700张JPG图像，其中530张是测试图像，其背景与训练集不同。
*Knifey类总共有1347个图像（测试集中有137个图像）。
*Spoony类总共有2208个图像（测试集中有242个图像）。
*Forky类总共有1145个图像（测试集中有151个图像）。

Knifey-Spoony数据集的下载

The archived tar-ball is automatically downloaded and extracted by using the [knifey.py](https://github.com/Hvass-Labs/TensorFlow-Tutorials/blob/master/knifey.py) module for Python.

Knifey-Spoony数据集的使用方法

DL之VGG16：基于VGG16(Keras)利用Knifey-Spoony数据集对网络架构FineTuning和迁移学习