Catch & Release, 2018-2020

add under-represented knowledge to machine learning

Catch & Release (C&R) is a collection of procedures that allow one to apply machine learning classification onto video. C&R uses a novel approach to label images from video with speech via speech to text, making use of the synchronicity between video and audio signals.

C&R’s goal is to facilitate the creation of under-represented knowledge in machine learning in general, and experimental datasets for neural network image classification in particular. C&R allows anyone with a mobile phone and a laptop to create viable datasets for image classification, and to train state of the art convolutional neural networks with these datasets.

GitHub repo