Competitions

As part of this course, students are required to participate in three Kaggle competitions. Each competition focuses on a different task in medical image analysis: classification, segmentation, and object detection. Sample notebooks will be provided as starting points, and students are required to beat the provided baseline.

Skin Cancer Classification

In the classification challenge, students will classify dermoscopic images into two categories: malignant (cancerous) and benign (non-cancerous). Images from the ISIC 2017 – Task 3 dataset will be used for this task. All solutions must be developed using deep learning models. The skin cancer classification competition will run from February 27 to March 20.

Kidney Tumor Segmentation

In the segmentation challenge, students will segment the kidney and tumors from CT scans. The 2019 Kidney Tumor Segmentation dataset will be used for this task. All solutions must be developed using deep learning models. The kidney segmentation competition will run from April 3 to April 24.

Liver Disease Detection

In the object detection challenge, students will detect and localize four liver disease patterns in histopathology images. The Liver Disease Detection in Histopathology Images dataset provided by Roboflow will be used in this competition. The goal of this challenge is to detect and localize four types of liver disease patterns: Ballooning, Fibrosis, Inflammation, and Steatosis. The Liver disease competition will run from May 8 to May 29.

Sample notebooks will be provided in the Code section, and detailed information about the datasets used in each Kaggle competition will be available in the Data section of the respective competition. Links to all competitions will be shared on Slack on their start dates.