Prototype to convert chyron text from docTR/Tesseract/LLaVA MMIF output into a name and list of attributes.
General user instructions for CLAMS apps are available at CLAMS Apps documentation.
This requires Python 3.8 or higher. For local installation of required Python modules, see requirements.txt.
For the full list of parameters, please refer to the app metadata from the CLAMS App Directory or the metadata.py file in this repository.
The app takes a MMIF file with TextDocument
annotations generated by an OCR app such as doctr-wrapper
or Tesseract.
The app outputs a TextDocument annotation corresponding to each
of the input annotations, containing the original text split into name-as-written and attributes fields
in an escaped JSON string; optionally, the output may also include a name-normalized field.
Each annotation also contains identifying information for the new annotation,
source annotation, and source VideoDocument.
For more details, see the output section of the app metadata.