AI-backed OCR in Healthcare
Daniela Gîfu
Abstract
As revolutionary technology, scanning technology - like OCR - knows an attracting increasing interest in the medical system. In fact, scanning technology, driven by big data and machine learning, helps to drive successful change processes in healthcare. Nowadays, the OCR systems based on the promise of artificial intelligence can contribute to a greater understanding of entire processes by automating the medical transcription and in general mining information (largely handwritten) clinical notes. The project aims to create an OCR application for offline character recognition using deep learning (a combination of RNN and CNN) that can be used in the medical sector. Concerning problems of medical care, large data such as dispersed dataset, weak consistency, low electronic degree, and low visualization degree, a set of innovative image recognition, dataset, arrangement, and storage projects were put forward. The proposed pilot method achieves results comparable to current SoTA.