PaintMaster - Reckognize paintings !
Your lack of knowledge in terms of classical art is terrible? When you see a painting, you can barely recognize the type of art, so of course not the painter?
Let's use PaintMaster, an app coded thanks to AI and Machine Learning to no longer embarass yourself! Paint Master would help you to finally recognize some paintings.
Principle:
The idea is to use ml5.js, a JS library created to make machine learning and AI coding more accessible to novices and students. It supports coders to learn about basic Artificial Intelligence methods and models.
To do so, we are going to use a fonctionnality of ml5.js called ImageClassifier. ml5.imageClassifier() is a method to create an object that classifies an image using a pre-trained model. It uses the MobileNet fonctionnality to determine with a certain probability from where our painting comes from.
We will have to train the algorithm thanks to Teachable Machine. Teachable Machine is a web-based tool that makes creating machine learning models fast, easy, and accessible to everyone. We will have to gather and group our examples into classes that we want the computer to learn. Then, we will train our model, and instantly test it out to see whether it can correctly classify new examples. Once the training is done, we would be ready to code our project.
First, you will have to enter some instructions in your glitch project, in the index.html file:
Then, it is important to have a part who can deal with the image of the painting that you have, so the AI can recognize with high probability the author and name of the painting.
Congratulations, you can now use Paint Master, for instance to identify what is and who painted amazing oeuvres, like this one:
Thanks to the work made with Teachable Machine, ml5.js and MobileNet fonctionnality, Paint Master would indicate you that the painting chosen is probably Ophelia, by John Everett Millais.
References:
TeachableMachine
ml5.js
GitHub
Scanner for IPhone
Machine Learning
Made by Alix Amerein