This is the finger-text detection and recognition module for the Smart Connect Worker project.
This repository is the Python implementation regarding the CRAFT-based text recognition and detection model as well as the finger detection model for implementing real-time human-machine interaction monitoring. This work is largely based on the text detection model of CRAFT.
The overall workflow of the finger-text detection and recognitionm module is as follows:
The weights used in this project are here: weight1 and weight2. You may also find the refiner model here. Put these weights into directory "/weights"
This is a demonstration of the finger and text recognition module, which automatically identifies the texts as well as the human's finger positions in real-time for the monitoring of human-machine interactions: