As with the emergence of Artificial Intelligence there is lot of growth of Artificial Intelligence in Medical Field Also. India is one of the developing countries in World with such a vast population. So, to treat everyone properly as well as Some people are not able to afford the proper treatment. And with normal middle-class mentality they are not supposed to call the doctor for their small diseases or symptoms that in future may lead to a big disease and can be a great threat. Nobody knows so precaution is better than Cure.
MediBot is a Platform where people are able to upload their CT-Scan reports and with the help of UNet architecture it would be able to predict whether a patient have Tumor or not. As well based on the provided prescription of the Patient it will extract the entities as in Symptom, Weights and medicine or treatment prescribed, so we are able to get an overview without going through the whole report. well as you are able to highlight the key points. With the chatbot assisting you for diseases.
In this I have finetuned gemini on the medical Dataset with the help of LORA and used it.
For the Tumor Detection I have used the UNet Architecture which is released in 2016 for the medical Dataset.
And for the name entity recognition I have used a Finetuned Model.
I received around 99% accuracy from the UNet Architecture.
And Chatbot also is working well
