Apples to apples

This project looks to implement a model that can recognise and classify images of fruit in real time.

The data used in this project can be found at the following link: https://drive.google.com/open?id=1-UJPLXEM7avZUFHLPRQtDogJPnGqNVKV

This data consists of:

• 'Fruit 360' dataset, which is available directly on Kaggle
• Flickr dataset - a set of data that I have collected, curated and labelled from Flickr
• Demonstration dataset - a set of high def images used for my capstone demo day
• 'FruitCategories.csv' A file the lists superclass and subclasses of fruit (e.g Apple -> Braeburn)

The project consists of 3 documents:

• Flickr Image scrape.ipynb - A jupyter notebook used with the Flickr API
• Capstone Project - Apples to apples.ipynb - A jupyter notebook detailing the EDA and modelling of the project
• Pi_DemonstrationTool.py - A .py file that was used to demo the tool

Dependencies

See attached document

Results