“AI-Powered Cancer Prediction and Image Segmentation System with Integrated Medical Chatbot”

AI-Powered Cancer Detection: Revolutionizing Early Diagnosis with Intelligent Systems

This project represents my journey in building a comprehensive AI-powered cancer prediction and diagnostic platform — integrating machine learning, deep learning, and medical AI to enhance early detection and assist healthcare professionals in making informed decisions.

What I Built

Breast Cancer Prediction

Developed a Support Vector Machine (SVM) model achieving 99% accuracy using 7 critical tumor features.

Created interactive visualizations to illustrate feature correlations with cancer risk.

Designed a CNN-based deep learning classifier for mammogram image analysis.

Integrated a bilingual (Arabic/English) AI medical chatbot with real-time speech recognition to assist patients interactively.

Brain Tumor Detection

Engineered a custom UNet architecture with Attention Gates for accurate MRI-based tumor segmentation.

Implemented confidence-based tumor detection and classification.

Built an auto-generated medical report system that provides risk-level assessment for doctors and patients.

Key Technical Highlights

Multi-model AI integration: SVM, CNN, UNet, Transformers

Explainable AI with feature importance and threshold-based risk scoring

Medical chatbot interface designed for patient-friendly interaction

Fully interactive Streamlit dashboard with dark mode visualization

Impact

This system bridges the gap between AI and oncology, offering:

Early and accurate detection of cancer types

Explainable and trustworthy AI tools for healthcare professionals

Accessible and inclusive interfaces that empower patients

Tech Stack

Python · TensorFlow · Scikit-learn · OpenCV · Plotly · HuggingFace Transformers · Streamlit