Keras-API

Keras + Flask + PreTrained Models

Flask configuration: https://pythonprogramming.net/basic-flask-website-tutorial/?completed=/practical-flask-introduction/

Set up the server at Digital Ocean

https://www.youtube.com/watch?v=qZNL4Ku1UQg&t=96s

Pre-trained models available

The models available are the ones that you can find through Keras. The following list contain the model used through the project:

InceptionResNetV2
InceptionV3
Xception
ResNet50
VGG16
VGG19
MobileNet
DenseNet

Example

To use a model you can use the kind of url: http://ip-server/ImageRecog/NameOfModel?image=LinkOfYourJPG

Example with the Pre-trained model ResNet50 with my own server: http://138.68.21.27/ImageRecog/ResNet50?image=https://www.uncompagnon.fr/upload/annonces/41972/middle/01-55637871_1.jpg

The output of the pretrained model ResNet50: