Recogniton of doodles from the ‘Quick, Draw!’ dataset.
The Objective of this project was to recreate a Neural Network using the Tensorflow Deep Learning Library to identify drawings at run-time. This model trains on ‘Quick, Draw!’’s dataset, this dataset is a collection of 50 million drawings across 345 categories, contributed by players of the game Quick, Draw!, for this project we will settle for 10 of them.
- Input data analysis and preprocessing;
- Building a model to classify drawings;
- Building a little software to draw;
- Process the drawing with the model;
The dataset will be in .npy format and shortened to 28x28 resolution to simplify it’s preprocessing.
A cat from the training dataset:
