This repository contains inference scripts for two different models designed for the detection of brain tumors: a classification model and a segmentation model.
The classification model is trained based on the notebook available at this link. It uses a Convolutional Neural Network (CNN) with the VGG16 architecture to classify whether a brain tumor is present in an image or not.
You can find the inference script for the classification model in inference.py and the Dockerfile in dockerfile.
The segmentation model is trained based on the notebook available at this link. It uses a U-Net with EfficientNetB7 for brain tumor segmentation.
You can find the inference script for the segmentation model in inference_rmi_tumor_segmentation.py and the Dockerfile in dockerfile.