Computer Vision is a field of Artificial Intelligence and Computer Science that deals with how computers can be made to gain an understanding of digital images and videos. With the advances in machine learning technology, innovations in Image Processing are emerging rapidly and being leveraged by businesses all around the world. Applications of Computer Vision include:
- Boredom killers such as your favorite Snapchat filter that use facial recognition and image processing to distort your face and/or overlay images
- Potential lifesaving uses such as Medical startups that claim they’ll soon be able to use computers to read X-rays, MRIs, and CT scans more rapidly and accurately than radiologists.
- And borderline unethical uses such as Facial Recognition in police body cams With so many emerging and interesting uses of Computer Vision, my goal is to compare different machine learning algorithms from scikit-learn and Keras tasked with classifying drawings made from Google's Quick Draw! dataset. By doing so I will showcase my abilities processing and analyzing the doodles through image classification and neural networks. This project is meant for companies that are interested in leveraging the Computer Vision methods used in my project in business.
Google has capitalized on the use of crowdsourcing to label over 50 million drawings with their online game "Quick Draw!". It gives its users 20 seconds to draw one of 345 different classes that range from an aircraft carrier to a zig-zag. Recently, they have open-sourced all their data and I will select 10 images to build my classifier. The data we used is a 28x28 grayscale bitmap in numpy .npy format of the simplified version of each drawing.