The current repository contains Davide Pollicino' Honours Project. The project (still under evolution), is the result of three research questions:

• What are possible approaches to feature enginnering for the implementation of a classifier able to distinguish flower variants incredibly similar between them, even at differnt life stages and growth locations;
• Which CNN architecture would offer best performances
• Is is possible to integrate a machine learning model within a django App, without that this model would first be deployed in a cloud service and exploses via endpoint?

The machine learning model, is the the integrated and used in a Django Web APP, where user are able to:

• Classify a flower
• Leave a feedback related to the prediction
• Save a prediction as favourite
• Register, Login, and gets the user's favourite position.

# create virtual environemnt python3 -m venv venv # install project requirements source venv/bin/activate pip3 install -r requirements.txt # run migrations python3 manage.py makemigrations python3 manage.py migrate # create superuser python3 manage.py creatersuper # run application python3 manage.py runserver 

(Note: manage.py may be located inside the webappClassifier folder :) )

• Coding style: Black
• Python Lint: Flake8, flake8-todos
• mypy (for english type checking)

# format folder black folder_name # Remote all white spaces from project files trim . # Format file to improve syntax in according to flake8 (yes -> 2 times --aggressive) autopep8 --in-place --aggressive --aggressive filename.py