Nameboard Detection and Text Recognition in OPDs of Public Hospitals

Unorganized and unmanageable manual handling of appointment or booking systems is a crucial issue in public institutions that provide facilities. This leads to unmanageable duties on staff members that make the entire system in mess. The scenario can be experienced in Outdoor Patients Departments (OPDs) of the Government Hospitals in Sri Lanka, leading to unexpected and unregulated crowds and queues of patients, due to the unmanageable appointment issuing especially during the rush hours of clinics and in pandemic situations. The current method of issuing appointment tokens lacks a systematic approach, often causing extended waiting times since doctor’s’ availability in respective rooms is not checked beforehand. In this work, we propose an automated system to address these challenges by identifying active rooms in the OPD section and detecting the name boards on doors and extract information to determine the availability of doctors. Door detection, open and closed-door identification, and name board detection are the three phases of the system. The key performance metrics include achieving an accuracy of $73.4 \%$ for automated door detection while name board detection and recognition resulted a success rate of $84 \%$ and $74 \%$, for closed doors and open doors respectively. Introducing a patient-doctor management database facilitates the systematic updating of room and doctor details, enhancing the efficiency of medical facilities utilization.