Image Labeling for Augmented Reality
DataForce annotates large dataset for an augmented reality application.
Our client, a global top five tech company, asked us to review 10,000 ordinary, everyday scenes from 500 homes, such as a bedroom or kitchen, and label polygonal bounding boxes applied to each object within that scene. Every object was to be labeled with their type and definition.
We utilized a tool that identified the placement of these polygonal bounding boxes, allowing annotators to categorize objects according to the taxonomy provided by our client. DataForce completed the project within six weeks at 98% accuracy.