Designed a machine learning model to predict the diseases from the images of chest X-Ray comprising of 14 diseases using novel approaches like mobile net, efficient net and try to build upon it using some new approaches like federated learning and wavelets based techniques.
Employed state of the art models, such as Inception, ResNet, MobileNet and EfficientNet on the NIH chest X-ray dataset comprising 100k+ images and 30k+ patients for multi-label classification of 14 diseases also implemented federated learning, boosting performance through decentralized client-server collaboration and designed a noise removal preprocessing unit utilizing Haar wavelets, targeting the LL component of images.
Designed a machine learning model to predict the diseases from the images of chest X-Ray comprising of 14 diseases using novel approaches like mobile net, efficient net and try to build upon it using some new approaches like federated learning and wavelets based techniques.