Virtual Radiologist

Identify Diseases with Ease

Comment

💡 Inspiration

With the use of artificial intelligence, Virtual Radiologist can identify pneumonia and brain tumors using X-rays. Many health problems go unnoticed because they are difficult to detect with an X-ray in their early stages. Detecting Pneumonia and Brain Tumors early on might save the lives of those who are suffering from them. This is where Deep-learning comes in. The depth of deep-learning networks distinguishes them from single-hidden-layer neural networks, which are more frequent.

❓ What it does

  • It aids in the early detection of potentially fatal illnesses.

  • Fill the void in the health-care industry due to a shortage of qualified personnel.

  • By just taking an X-ray, a regular checkup will be available. Reduce hospital overpopulation, particularly during times of crisis.

  • In comparison to other options, it is less costly. Reduce the number of cases of illness carelessness caused by poor imaging or cases handled by medical personnel. Suspected cases are identified more quickly.

*The categorization of items such as lesions into categories such as normal or aberrant, lesion or non-lesion, is done using image analysis.

🔨 How we built it

  • In the frontend we have used HTML, CSS, JavaScript and for UI I have used bootstrap

  • At the backend, We have used NodeJS

  • We have built the machine learning modal using TensorFlow and CNN(convolution neural network)

  • We have used Google Colab to train our model which took so much time and was the hardest time

🏃‍♂️ Challenges we ran into

The key issue was the length of time required for such projects. Training a model takes roughly 8 hours on average, and training two of them was a difficult task. It was also a difficult effort to extract such large models so that they could be used on the web. After the model was extracted and used on the web, the next difficult step was to pre-process the input image to fit it into the model for prediction, which no web programming language could do, so we had to find another solution.

It was difficult for us to create noises and animations in the frontend to make the UI engaging. It was also difficult for us to integrate the frontend and backend.

🏆 Accomplishments that we're proud of

We are proud on ourselves that we completed this project within the time limit

📚 What we learned

  • We learned how to use bootstrap to design UI in website.

  • We leaned how to build disease prediction modal using TensorFlow

  • We also learned about CNN(convolution neutral network)

  • Furthermore, We also learned how to use Node.js at backend

⏭ What's next for Virtual Radiologist

In the future, we want to add more diseases' prediction modal in the web application to the web application, so that doctors can use it to predict diseases and get more result.

Built With

Share this project:

Updates