Historical archives often have to deal with a huge amount of photographs and other historical images. Once they are digitalized, they need to get categorized: What is pictured on them?

This application is an experimental approach to use automated image-recognition for this task.

This first version runs with Google Vision. A further goal is to integrate other picture-recognition-APIs as well to get better results (e.g. Microsoft Computer Vision, IBM Visual Recognition, etc.).

• Source-Code: https://github.com/zack17/klio-picture-labeler
• Demo-Page: https://zack17.github.io/klio-picture-labeler/

The used demo photos were taken from:

• Indien-Bilder der Handelsfirma Gebrüder Volkart, Stadtarchiv Winterthur, Link
• Bildarchiv des Lebensmittelverein Zürich (LVZ), Sozialarchiv Zürich, Link
• Fotosammlung: Wandervogel - Schweizerischer Bund für alkoholfreie Jugendwanderungen, Sozialarchiv Zürich, Link

All licensed under CC-by

• Micha Reiser | GitHub
• Dominic Studer (doemee)