MLViewer

User interface for handwritten digits recognition and VGG16 neural networks using Qt Creator and TensorFlow.

Getting Started

Follow these instructions to get a copy of the project and make it run on your local machine (only tested with Ubuntu 18.04).

Prerequisites

You need to have these tools installed on your Ubuntu machine:

• Python 3 (pre-installed on Ubuntu 16.04 & 18.04)
• TensorFlow
• Qt Creator

Settings

Once you have cloned and opened this project with Qt Creator, you have to disable the Shadow Build in the Projects tab. Before running the app, you also have to export a handwritten digits recognition model using the MNISTER repository. A VGG16 trained neural network can be found using this repository. Then, you have to make sure you have the following directory tree:

./WORKSPACE
	|-- icons
	|	|-- chip_256.png
	|	|-- image.png
	|	|-- number-blocks.png
	|-- MNIST
	|	|-- own_0.png
	|	|-- own_1.png
	|	|-- own_2.png
	|	|-- own_3.png
	|	|-- own_4.png
	|	|-- SAVED_MODEL
	|   	|     	|-- 1
	|   	|	|   |-- variables
	|   	|	|   |    |-- variables.data
	|   	|	|   |    |-- variables.index
	|   	|	|   |-- saved_model.pb
	|   	|	|-- graph_for_model_1_xxxxxxxx
	|   	|	|   |-- events.out.tfevents.xxxxxxxxxx.user
	|-- scripts
	|	|-- base.py
	|	|-- init_test.py
	|	|-- mnist.py
	|	|-- mnist_load.py
	|	|-- mnist_preproc.py
	|	|-- synset.txt
	|	|-- utils.py
	|	|-- vgg16.py
	|-- VGG16
	|	|-- cat.jpg
	|	|-- weasel.png
	|	|-- vgg16.pb
	|-- logwindow.cpp/.h
	|-- main.cpp
	|-- mlviewer.cpp/.h
	|-- MLViewer
	|-- MLViewer.pro
	|-- README.md
	|-- stylesheet.qss	

Running the app

MNIST

VGG16

TODO

• Load every type of model: allow user to specify input and output layers name of trained model
• Zoom on small pictures
• Reset only the output for MNIST

Icons

Credits: Icons made by Freepik from www.flaticon.com

Authors

• Pierre OLIVIER - polivier2