基于深度学习的13种通用图像分类模型及其实现

image classification with deep learning model： VGG16、VGG19、InceptionV3、Xception、MobileNet、AlexNet、LeNet、ZF_Net、ResNet18、ResNet34、ResNet50、ResNet_101、ResNet_152

------------------------------------------------------------------------------------------------------

-------------------------------------------------------------------------------------------------------

项目集成了 VGG16、VGG19、InceptionV3、Xception、MobileNet、AlexNet、LeNet、ZF_Net、ResNet18、ResNet34、ResNet50、ResNet_101、ResNet_152等13种图像分类模型作图像分类，依据测试结果来看，残差网络的分类准确率最高，分类效果最好

项目地址 https://github.com/tslgithub/image-classification-VGG-InceptionV3-Xception-MobileNet-AlexNet-LeNet-ZF_Net-ResNet

the project apply the following models:

• VGG16
• VGG19
• InceptionV3
• Xception
• MobileNet
• AlexNet
• LeNet
• ZF_Net
• ResNet18
• ResNet34
• ResNet50
• ResNet_101
• ResNet_152

your train or test datasets folder should be:

/dataset/train/

• 1
  cat.jpg,
  cat2.jpg,
  cat3.jpg,
  cat4.jpg,
  cat5.jpg,
  cat100.jpg
  cat1000.jpg
• 2
  cat.jpg,
  cat2.jpg,
  cat3.jpg,
  cat4.jpg,
  cat5.jpg,
  cat100.jpg
  cat1000.jpg
• 3
  cat.jpg,
  cat2.jpg,
  cat3.jpg,
  cat4.jpg,
  cat5.jpg,
  cat100.jpg
  cat1000.jpg
• 4
  cat.jpg,
  cat2.jpg,
  cat3.jpg,
  cat4.jpg,
  cat5.jpg,
  cat100.jpg
  cat1000.jpg

/dataset/test/

• 1
  cat.jpg,
  cat2.jpg,
  cat3.jpg,
  cat4.jpg,
  cat5.jpg,
  cat100.jpg
  cat1000.jpg
• 2
  cat.jpg,
  cat2.jpg,
  cat3.jpg,
  cat4.jpg,
  cat5.jpg,
  cat100.jpg
  cat1000.jpg
• 3
  cat.jpg,
  cat2.jpg,
  cat3.jpg,
  cat4.jpg,
  cat5.jpg,
  cat100.jpg
  cat1000.jpg
• 4
  cat.jpg,
  cat2.jpg,
  cat3.jpg,
  cat4.jpg,
  cat5.jpg,
  cat100.jpg
  cat1000.jpg

1,2,3,4 is classes name or folder name,whose path is

"training data set folder is:"

/dataset/train/1/cat*.jpg,

/dataset/train/2/cat*.jpg,

/dataset/train/3/cat*.jpg,

/dataset/train/4/cat*.jpg,

"testing data set folder is:"

/dataset/test/1/cat*.jpg,

/dataset/test/2/cat*.jpg,

/dataset/test/3/cat*.jpg,

/dataset/test/4/cat*.jpg,

• Attention: classes name ‘1,2,3,4’ or folder name must be number, not string

environment

my environment is based on ubuntu16、cuda8、tensorflow_gpu1.4, all package needed can be installed with ‘pip3 install package_name’, and you can test which package is missed by run ‘python train.py’,then pip install the missed package

train and predict your model

train model: python train.py

predict model: python predict

Any Questions???

Author email: [email protected]