Research

Medical Image Classification Study

A cautious research prototype exploring CNN-based classification workflows, evaluation metrics, and model limitations.

ResearchResearchMLAI
Screenshot placeholder: training-curves
Screenshot placeholder: confusion-matrix

Overview

This project studies image classification as a learning exercise. It avoids medical claims and focuses on responsible evaluation, dataset limitations, and reproducible training.

Evaluation

The report includes accuracy, precision, recall, confusion matrices, and notes about class imbalance.