Wrapper for EasyOCR end-to-end text detection and recognition.
The wrapper takes a VideoDocument
with SWT
TimeFrame
annotations and
returns BoundingBox
and
TextDocument
aligned
with TimePoint
annotations from the middle frame of each TimeFrame.
EasyOCR seems to handle reading order, but there isn't much documentation regarding the implementation. Regardless,
the BoundingBoxes
and TextDocuments
are ordered within each TimePoint
so reading order can be inferred from the
annotation ids.
Here's the framework from the EasyOCR repo:
General user instructions for CLAMS apps is available at CLAMS Apps documentation.
- Requires mmif-python[cv] for the
VideoDocument
helper functions - Requires GPU to run at a reasonable speed
For the full list of parameters, please refer to the app metadata from CLAMS App Directory or metadata.py
file in this repository.