This is the software implementation for the article "MobileNetV2 Ensemble for Cervical Precancerous Lesions Classification": https://doi.org/10.3390/pr8050595
The article is free. It is a computer-aided diagnosis system to help the doctors in colposcopy (a medical investigation for preventing cervical cancer). I would highly recommend reading the article in order to understand the implementation and the related experiments.
The authors propose an ensemble of Mobilenetv2 convolutional neural networks. It classifies colposcopy images according to the severity of the cervical lesions.
Everything has been developed as Jupyter notebooks using Python and PyTorch frameworks. The files having tar extensions represent neural network checkpoints. Those can be loaded in other applications. Please check the source code for examples of how to load the checkpoints.
If you are not familiar with Jupyter notebooks, I recommend starting with Google Colab, a free environment for experimentation. There, you can run Jupyter notebooks on a virtual machine hosted by Google.