Emory BMI GSoC Project Ideas

Niffler: A DICOM Framework for Machine Learning and Processing Pipelines.

Deep Learning approaches in the detection of pulmonary disorders: COVID19, Tuberculosis, Bacterial, and Viral Pneumonia, Healthy/Normal using 17500 non-augmented X-ray images. 5 class classification performed using different pre-trained models like DenseNet201, Xception, Inception, and many more reaching near 99% accuracy.

Locate basic landmarks on cephalograms with AI (Pytorch)

The project is a collaboration with David Loaiza ( 4th Yr Radiologist) from Mexico at Cardiology national institute "Ignacio Chavez". The aim is to estimate the bone age from the left hand radiographs. The model will be trained on a RSNA Pediatric Bone Age Challenge (2017) public dataset and evaluated on private dataset obtained from the hospital.

Use an android phone to easily scroll through radiology image stacks and more

A lightweight and efficient Dicom server for receiving and storing radiological images and radiation dose structured reports