Accelerated Seven Segment Optical Character Recognition Algorithm

We present an accelerated optical character recognition approach of seven-segment display digits. We focus on digital medical devices, which are used by aged people with possible disabilities, although numerous practical applications are available for this type of digit recognition. Using the proposed system, patients are expected to scan a medical instrument with a smartphone to automatically extract measurements related to their health. Such a system can make the procedure easier for the patient and does not require specific knowledge of what each number represents. Although, similar research is existent, we present a highly parallelizable algorithm, along with it's implementation on an FPGA host, to accelerate the recognition process. Experiment results on FPGA and ARM platforms illustrate that our enhanced multithreaded procedure of applying optical character recognition achieves 96.22% accuracy and features high scalability with a factor of 0.9.