Doctor Handwritten Recognition and Alternate Medicine Recommendation using Deep Learning

Doctors’ illegible handwriting in prescriptions cause trouble to both patients and pharmacists in many ways. Due to their handwriting pharmacists may not understand the medicine names and there is a chance that they may misunderstand the medicine names. If those misunderstood medicines are given to the patients rather than the prescribed ones, then that could cause them some other harm, there are chances that even that patients may die. Main problem here is that the prescription cannot be read. So, this proposed work would work as a solution for that. This proposed work can read the medical prescription when the image if the prescription is given character by character. Then along with that this proposed model could also recommend the alternate low-cost medicine. Many people do not know the actual alternate medicine they can cure their disease at a low cost. But this is not acceptable for all cases though. This will be very helpful for those people who are not able to afford the medicine cost. For medical prescription recognition a model which is a combination of AlexNet and SVM Classifier is used. AlexNet, which is a transfer learning model, is specifically used for feature extraction and SVM Classifier is used for classification. Through this process a remarkable accuracy of 98% has been achieved. This process starts with taking images as input and then the features extracted from the images with transfer learning model, then classification done using the SVM Classifier. That gives the recognized medicine name, using that name an alternate low-cost medicine will be given.