Skip to content

clamsproject/app-parseqocr-wrapper

Repository files navigation

PARSeq Scene Text Recognition Wrapper

Description

This app wraps the PARSeq scene text recognition engine into a CLAMS app.

User instruction

General user instructions for CLAMS apps is available at CLAMS Apps documentation.

System requirments

To install this app locally, one has to install the underlying OCR engine parseq. Specifically, the version of parseq used by this app is at commit bc8d95c. In most cases (on a non-GPU computer with an Intel or AMC (amd64 architecture) CPU), the following command should work:

# set up your python environment first before going into the following
# e.g. `conda create blah blah`
# now, from the directory where you want to download the source code
git clone https://github.com/baudm/parseq.git
cd parseq
git checkout bc8d95cda4666d32fa53daf2ea97ff712b71e7c7
pip install -r requirements/core.txt -e . 
# let's confirm it's installed by running the following command
## get out of the cloned parseq source directory
cd ~
## try import parseq into python 
python -c "import strhub"

If you see an ImportError from the above, the installation went wrong. Please visit the parseq repository for more detailed installation instructions, including instructions for other computing platforms such as GPU-enabled computers, ARM-based computers, etc.

Configurable runtime parameter

For the full list of parameters, please refer to the app metadata from CLAMS App Directory or metadata.py file in this repository.