SymptoScan

SymptoScan

A health clan, from SymptoScan for your Lifespan!

The problem SymptoScan solves

Whether you're experiencing a mild fever, persistent cough, or simply have questions about your well-being, We the team JDSC! have built SymptoScan powered by ML algorithms for you. Through thoroughly tested data and confidential analysis on your fingertips, we empower you to:

Receive early disease assessments: Early diagnosis of critical conditions such as stroke, dysrhythmia, Parkinson's, and pneumonia is challenging, often leading to compromised prognoses. Get insights into potential causes of your symptoms, helping you understand what might be going on.

Gain clarity on your symptoms: Describe your experiences and symptoms and our state-of-the-art AI technology will expertly analyze them against a vast database of medical knowledge including further insights into diseases like Pneumonia, Stroke, Parkinsons' and many more.

Empower informed decision-making: Access trusted health information, curated by medical professionals, to feel confident in managing your health journey including all the medical information.

Connect with the right resources: Based on your assessment, SymptoScan will recommend reputable healthcare providers, specialists, and self-care options tailored to your needs as well as all the initial precautions that can be taken.

Challenges we ran into

Getting reliable datasets in the medical sector is always a challenge as the data has to be official and the models cannot be trained on any kind of misleading data. We aimed to utilize multiple sources of data including text, audio, and video further making them seamless for the users' benefit was challenging. Finally, achieving decent accuracy scores worthy of being termed initial medical diagnosis.
(Also official video submission has been made on an accessible GOOGLE DRIVE and the link of the drive is the LAST IN THE LINKS SECTION)

Tracks Applied (1)

Core Machine Learning

Our project uses multiple Core-ML concepts including CNNs with an accuracy of 94.8%, XGB Classifiers with an accuracy of...Read More

Discussion