This approach has resulted in recognition accuracies above 93%, offering a promising solution for digitising handwritten prescriptions and reducing transcription errors in healthcare [4].