AutoTag: automated metadata tagging for film post-production

Marcelo Sandoval-Castañeda, Scandar Copti, Dennis Shasha

Research output: Contribution to journalArticlepeer-review


Film post-production can be time- and money-inefficient. The reason is that a lot of the work involves a person or group of people, called metadata taggers, going through each individual piece of media and marking it up with relevant tags, such as the scene number, transcripts, and the type of shot for video footage. Such a task is particularly time-consuming for films with high shooting ratios (i.e., footage shot/footage shown). AutoTag automates much of the tagging process across 16 languages, saving both time and money. We describe the algorithms and implementation of AutoTag and report on some case studies.

Original languageEnglish (US)
Pages (from-to)6731-6753
Number of pages23
JournalMultimedia Tools and Applications
Issue number3
StatePublished - Jan 2024


  • Film editing
  • Google speech-to-text
  • Image classification
  • Post-production
  • Speech recognition

ASJC Scopus subject areas

  • Software
  • Media Technology
  • Hardware and Architecture
  • Computer Networks and Communications


Dive into the research topics of 'AutoTag: automated metadata tagging for film post-production'. Together they form a unique fingerprint.

Cite this