A New Approach to Information Retrieval based on Keyword Spotting from Handwritten Medical Prescriptions

Author(s)

Arghya Mukhejee, Arunit Halder, Subhrapratim Nath, Subir Kumar Sarkar

Author affiliation

1Departments of CSE, Meghnad Saha Institute of Technology, Kolkata 700150, India 
2Electronics & Telecommunication Engineering, Jadavpur University, Kolkata 700032, India

Abstract

Handwriting recognition of medical prescriptions has been a challenging problem over the recent years with constant research in providing possible accurate solutions. Indecipherable handwritten prescription and inefficacy of Pharmacist to understand the medical prescription can lead to serious and harmful effect to the patients. In this paper, a document retrieval approach based on word spotting for medical prescriptions is proposed. The proposed approach involves extracting the domain specific knowledge of doctors by recognising printed text on letterhead part of the prescription in order to narrow down the search operation for word spotting in handwritten portion of the prescription. Handwritten word spotting performs on offline text portion using Hidden Markov Model. Experiment is performs on 500 prescriptions with a large heterogeneous medical data in varied environment to evaluate the effectiveness of the algorithm and thereby reveal the cognitive interpretation of Handwritten Prescriptions.

Keywords

Handwritten recognition; Medical prescription; Word spotting; Line segmentation