Machine Learning approach for the analysis of lung cancer data to aid the radiologists at global level
The core of the project is a Convolutional Neural Network (CNN) model, which is highly effective for image-based analysis of lung cancer. In addition to disease classification, it will also give the information about the location of lung cancer. Then this information would be transferred to the Natural Language Processing (NLP) model, which would be generating diagnostic reports like radiologists. In addition to this the medical images would also be used to create the 3D models for the better understanding. Hence, this tool would be significantly reducing the burden as well as the time from 45 minutes to 5 minutes.
In conclusion, this project represents a transformative approach to healthcare diagnostics in India, addressing both the shortage of radiologists and the need for rapid, reliable diagnoses. By providing a comprehensive solution that includes disease detection, localization, and report generation, it aims to make high-quality healthcare more accessible and efficient across India’s diverse healthcare landscape.
Successful model will be presented to the global scientific community for further enhancements and dissemination.
1. Address Radiologist Shortage: Mitigate India’s shortage of radiologists by providing an automated tool for lung disease diagnosis, reducing reliance on human radiologists.
2. Rapid and Accurate Diagnostics: Deliver fast and reliable diagnoses of lung diseases, including tuberculosis, lung cancer, and pneumonia, ensuring timely treatment and improved patient outcomes.
3. Disease Classification Using CNNs: Employ Convolutional Neural Networks (CNNs) to accurately classify lung diseases from medical images.
4. Precise Disease Localization: Integrate 3D modelling for accurate localization of disease-affected areas within the lungs.
5. NLP for Report Generation: Use NLP to produce detailed diagnostic reports within 3 to 5 minutes, significantly reducing report preparation time compared to traditional methods.
Prof. Tiratha Raj Singh as PI and Mr. Devansh Kalia as student.