Code for our Research paper(Under Review) project called "Averting from CNNs for medical image classification".
This paper attempts to apply newer approaches that do not use CNNs conventionally to the evolving field of medical image classification. While analyzing, firstly, an all MLP architecture MLP-Mixer and secondly, the inverted convolutional kernels coined as Involution with our baseline ResNets, both models yield comparable results in detecting Covid19 and pneumonia using Chest X-ray images. On top of that, merging Involution kernels into ResNet architectures can produce promising performance while training on roughly 40% fewer parameters. This paper further compares these two architectures with various CNN-based models. We hope this research further helps the research community to utilize the capabilities of these newly introduced architectures in the medical field.
(Note: Vgg19 is treated as an outlier due to its high parameter size.)