Overview

Melanoma is the most serious type of skin cancer, and early detection is crucial for successful treatment. Automated detection from dermatoscopic images is a classic problem in medical image analysis. In this competition, your task is to build a model that determines whether a skin lesion is benign or malignant (melanoma) based on its visual features.

You will work with a labeled dataset of high-resolution dermatoscopic images, giving you an opportunity to practice essential Computer Vision (CV) skills such as image preprocessing, data augmentation, and deep learning. This challenge is beginner-friendly and walks you through the full workflow of building an image-based machine learning model — from exploring medical data to generating accurate predictions.

Practice Skills

In this challenge, you will gain hands-on experience with:

• Python
• Image Processing and Computer Vision (CV)
• Deep Learning
• Medical Data Classification

Evaluation

Goal

Train a classification model that predicts whether each skin lesion in the test set is Benign or Malignant (Melanoma).

Metric

Submissions are evaluated on Accuracy:

$\text{Accuracy} = \frac{\text{Correct Predictions}}{\text{Total Predictions}}$

Submission Format

See the submission guide for full instructions.

The submission file has two fields:

• id: The image filename.
• label: The predicted label for the corresponding image (0 = Benign, 1 = Melanoma).

id	label
9c9a204a57f34a71b9454fef633f7038.jpg	0
cec04296a9bf4b969a8770f3ee2b439f.jpg	1

Submission Types

The challenge accepts two types of submissions:

Public Submission

Use the public test set (provided in the Data section) to generate your submission file.

Private Submission

• You must submit your model.
• It will be evaluated on a private test set.
• Important: Your code must recursively scan the entire data directory (see the Code section for implementation details).