DIAGNijmegen / bodyct-luna22-ismi-algorithm-baseline

Repo for creating the baseline algorithm container

Algorithm for classifying lung nodules in chest CT

The source code for the algorithm container for classifying 3D blocks of lung nodules from chest CT (shaped [128, 128, 64] in x, y, z), generated with evalutils version 0.3.1.

This algorithm contains the inference scripts for VGG16 models that estimate the malignancy risk and predict the nodule type of lung nodules.

This GitHub repository is linked to the baseline algorithm for the LUNA22-ISMI educational challenge, a challenge for the Intelligent Systems in Medical Imaging course at the Radboud University.

Interfaces

Inputs

• CT image (/inputs/images/ct/<uuid>.mha)

Outputs

• Lung nodule malignancy risk (/outputs/lung-nodule-malignancy-risk.json)
• Lung nodule type (/outputs/lung-nodule-type.json)