Carnegie Mellon University
Browse
2020.10.8_CAMPI_Whitepaper.pdf (6.64 MB)

CAMPI: Computer-Aided Metadata Generation for Photo archives Initiative

Download (6.64 MB)
Version 2 2020-10-09, 14:27
Version 1 2020-10-08, 22:48
preprint
posted on 2020-10-09, 14:27 authored by Matthew LincolnMatthew Lincoln, Julia CorrinJulia Corrin, Emily DavisEmily Davis, Scott B. WeingartScott B. Weingart
In the summer of 2020, CAMPI (the Computer-Aided Metadata Generation for Photoarchvies Initiative) assessed the use of computer vision in assisting in the arduous task of processing digital photograph collections. The prototype built to find duplicates and tag photos depicting similar scenes in Carnegie Mellon University Archives’ General Photography Collection was successful, proving such an interface would be both feasible and useful for cultural heritage institutions with large visual collections.

This whitepaper reports on the background for this project, the specific tasks, methods, and results from our software prototype, and a summary of future directions and needs for both computer vision as well as interfaces for photo archive description. We also offer a high-level technical roadmap as well as specific implementation details and reference code for our prototype.

History

Date

2020-10-08

Usage metrics

    Categories

    Keywords

    photo archivesarchivesvisual similaritycomputer visioncollection management system

    Licence

    CC BY-NC-ND 4.0

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC