Vision data enrichment and annotation platform

Abstract

A platform for data collection and annotation is disclosed. The systems and methods herein describe a system that can receive vision data from a plurality of heterogenous sources and store this vision data in a common datastore. The vision data is then processed (for example, prioritized and annotated) at common applications, while security and access rights features are maintained. Machine learning models for annotation are trained in such a way as to generate an efficient and accurate model, while ensuring that manual annotation costs remain below a set budget.