Book Annotation Classification

About the Project

The purpose of the Zooniverse Book Annotation Classification project is to generate training data for the Book Annotation Classification Engine project, a machine-learning experiment to detect and classify annotations in digitized printed books.

Annotations and other marks are traces left behind by previous readers and owners. Through these markings, we can better understand how readers in the past have used their books or interpreted their contents.

In Phase I, we are collecting data to:

1. Identify pages that contain annotations
2. Delineate page regions with the annotations
3. Classify various annotation types

Annotation Detection / Classification

The data collected will serve as training data for the machine-learning experiments to develop a tool that researchers can use to discover annotated pages in IIIF-hosted digital collections. All resulting data will be made publicly available via our UCLA Collections Lab GitHub space.