Model for COVID detection in X-Rays Further a Flask based platform is finished for easy inference on images. Developed @BISAG
Augemented + Segmented + Equalized : https://drive.google.com/drive/folders/1VP-g1lDvGn-WytEFYnvoBe1oqsMtkV2n?usp=sharing
Original Data : https://github.com/ieee8023/covid-chestxray-dataset
Input Image -> Augmentation -> Segmentation -> Mask Cropping -> VGG16 -> Diagnosis
CAM analysis on non-segmented images, model learns irrelevant features. Overfitting to other source semantics.
CAM analysis on segmented images, model learns important features which are further verified by professional diagnosis of the same image.