Skin Cancer Detection

Overview

With advancements in technology and especially AI/ML, we wanted to showcase our skills and dive deep into big data related to healthcare. Utilizing our knowledge in machine learning, we explored the ISIC 2024 Skin Cancer Detection Challenge — developing models to classify lesions as benign or malignant, analyzing metadata and image features from 3D Total Body Photography (TBP), and aiming to contribute toward more accurate and accessible early skin cancer diagnosis.

The Dataset

With over 400,000 data points and 55 features, we explored the data, cleaned up the data, and (while trying our very best) tuned our ML models to achieving predicting the probability of a lesion is benign or malignant.

Challenges we ran into

Being our first (in-person) hackathon, it took some time to come up with ideas and keep adjusting on the fly. With just being a group of two the work we wanted to do was ambitious and came with lots tasks to cover. However we we're able to do as much as we wanted to in under 24 hours and can't wait to work more on this project later down the line.

Built With

• matplotlib
• numpy
• pandas
• python
• scikit-learn
• tensorflow