4th year project
Includes: Keras CNN built with a tensorflow backend to identify football crests. Complete with a coreml conversion to be used in an ios app.
Script to download all images of all the premier league players and organise them into folders based on teams.
Django server app providing RestFul service to the ios app on https://github.com/seanjohn85/Elite-Premier-League. The Django app scraps the internet and uses a premier league API to gather information on the efl football teams and stores it in an SQLite DB. It also makes predictions on the teams next fixture using a Lamba function with Poisson Distribution algorithm(https://en.wikipedia.org/wiki/Poisson_distribution)
Everything is developed in Python 2.7 and it is recommended that a new virtual environment is set up for each fold using a system such as Anaconda.
All intructions are based on pip install.
Open a terminal in a new python 2.7 environment and enter the following commands
sudo easy_install pip pip install Keras==2.0.6 pip install tensorflow == 1.2.1 pip install -U coremltools pip install -U scikit-learn pip install pillow pip install h5py pip install matplotlib pip install numpy
**** it is important the version numbers are correct as they are required for coreml
The CNN model is located at https://github.com/seanjohn85/Premier-League/tree/Downloaded-Images/Python%20model/ConVnet
model/ConVnet in this repo
Open this folder in an editor such as spider
ensure that the file explorer is pointing to the location to where the images are that are been feed into the CNN model (in this case its the dataset folder)
run the script CnnMain.py
When the script completes it should produce a new core ml file that can be added to an iOS project
In order to get the image resources for the server set up a new environment
pip install requests pip install beautifulsoup4 pip install urllib
To download the images run the save images script.py in the Scrapping Folder in this repo. Then all the images are downloaded they will be stored in 20 folders based on their theme code
Move all the folders to the following folder Server App - epl - static - images
set up a new enviroment and install the following
pip install Django==1.11 pip install djangorestframework pip install markdown pip install django-filter pip install simplejson pip install requests pip install beautifulsoup4 pip install unirest pip install urllib pip install urllib2
Open a terminal window and navigate to the efl folder in server app folder in this repo. enter the following python manage.py makemigrations rest python manage.py migrate now enter ifconfig to get your current ip
python manage.py runserver curentIP:8080
the server will now start on this ip and can be accessed by the ios app