Diabetes is a disease that can kill you. Due to the lack of target assessment equipment for automated detection, almost 84 percent of diabetics are unaware that they have the condition. Diabetic Retinopathy is an eye disease that can cause diabetics to lose their vision and become blind. Diabetic Retinopathy should be identified quickly and correctly, yet clinician variability exacerbates the cutting-edge DR treatment's unfavourable outcomes.
Visara is a web-app that uses a range of cutting-edge models to transform the profession of ophthalmology. Traveling to gyms/yoga centres is neither safe nor viable due to the current pandemic, which brings our concept to light. Using a Yoga Bot that recognises the user's posture while performing the Asana, the web-app would not only predict the DR and the level of blindness, but it would also propose Yoga poses that have been shown to be useful in the treatment of Diabetic Retinopathy.
According to studies, simple report organisation leads to greater treatment efficiency. Furthermore, if a medical text report is unstructured, doctors are more likely to overlook vital information from their patients, therefore this summarizer ensures that no information is lost in the process of interacting with a doctor. As a result, a report is prepared based on the diagnosis results, which may be accessible by both the doctor and the patient for evaluation.
Our Visara project aims to provide a solution that will enable people to take the required precautions against one of the most prevalent problems that diabetics experience, eyesight loss. The project's main goal is to help patients and doctors use resources and funds more efficiently. The most appealing aspect of this project is that it attempts to automate practically everything, from the detection of Diabetic Retinopathy (DR) severity prediction and blindness time prediction to the creation of a report summarizer and the handing over of the details to our YogaBot.
Our YogaBot asks new customers a few general questions about their medical history before recommending Yoga asanas to improve their health and immunity. We maintain track of registered users' prior asanas and their progress in order to recommend new routines. Our bot then functions as a personal trainer, recording the body structure and alerting us if it is maintained properly.
Our YogaBot is a health analysis function that recommends and teaches Asanas to improve the user's health structure. It functions as a personal trainer, ensuring that the user's posture is correct while completing the Asanas.
The three primary aspects of Diabetic Retinopathy (DR) are severity prediction, blindness time prediction, and a report summarizer. Each of these attributes is represented by a set of records, to which doctors can add a variety of records by picking one of the three options. They can retrieve an output in seconds after filling out the appropriate information for each entry.
-
For the DR severity prediction part of our web app, we will use an image of the patient's retina to determine the stage of DR. Our approach rates the retina on a scale of 0 to 4, with 0 indicating no DR and 4 indicating severe DR.
-
The following component of our online programme calculates the likelihood of a patient becoming blind as a result of DR over the course of 70 months. A doctor can submit information on a patient's demographics and therapy. We want to generate a graph displaying the percent likelihood of going blind over the course of months using a machine learning model. This is beneficial to both the doctor and the patient because they may quickly determine whether or not they require therapy without having their retinas scanned.
-
Finally, we have a report summarizer that allows the doctor to quickly see a summary of the patient's status. Patients can also get a more condensed and organised version of their doctor's report.
Because there is no objective diagnostic technique for automated screening, almost 84 percent of patients are unaware they have diabetes. Detecting DR is now a time-consuming process that necessitates the evaluation of the retina by a skilled doctor. Furthermore, ophthalmologists' diagnoses of patients can be quite uneven, and our severity predictor allows an objective and speedy assessment of DR. Furthermore, we provide each client with a personalised set of yoga asanas as well as a personal posture trainer, all without the risk of Covid transmission.
- Use a messaging tool to connect the patient with the doctor.
- Diet Planner based on the patient's data.
- Using IoT devices to keep track of patient fitness and food history stats.
Our online programme predicts if a person will develop Diabetic Retinopathy and, if so, for how long, as well as proposing yoga asanas and connecting your reports with your selected doctor to help you naturally treat the disease.
Diabetic retinopathy is the biggest cause of blindness in working-age adults aged 20 to 74, and the overall covid and social distance environment promotes a healthcare model that is digital at its core, like ours.
At this stage, in addition to various digital marketing strategies, our first goal would be to establish a good repo with the doctors who treat such patients by automating the detection and cure process, which will be the main point to attract these doctors to recommend or treat their patients through our platform.
Some traditional doctors may make our platform less appealing in the market, but we are optimistic that technology will overcome stereotyping.
Since recent events have made it impossible for us to leave our homes, now is the greatest opportunity to gain control of a market that affects roughly 93 million adults worldwide. According to a recent survey, about 63 percent of Indians are unaware that diabetes has a negative impact on the eye, in addition to other body parts. After their vision was compromised, 92 percent of diabetics had their retinas examined. So, similar to cancer, where a person can save his or her vision and live a healthy life, the key goal would be to inform individuals at an early stage so that they can take the necessary procedures.
Ways to monetize our system include: - Creating a customer-based service that provides them with insights to keep them engaged with the platform.
-
Provide a service in which a separate database is provided to a specific clinic so that it may monitor the situation of its clients, transfer information quickly, and save time for the doctor by avoiding unnecessary patient visits.
-
As we refine and improve our product, we will be able to offer a distinct health consulting service to assist clients in staying fit by practising yoga asanas.
- In case link doesn't work, follow these steps:
- Upload
coxnetTR.pklandcoxnetUT.pklmodel files to your drive - Import the model files and execute the above colab file
- Upload
- Colaboratory file for DR prediction
- Diabetic-Retinopathy-Prediction(NGROK)
- Then follow these steps:
- Save the model files
d5 weights.h5to your hard drive - Run the above colab file after importing these model files.