🧠 Breast Cancer Treatment Progress Prediction Using SSIM

This project focuses on quantifying breast cancer treatment progress by comparing pre-treatment and post-treatment mammogram images using:

• CNN-based classification
• Structural Similarity Index (SSIM) for visual progress measurement
• Automated medical report generation

⚠️ Source code is intentionally private due to medical data sensitivity, institutional IP policies, and patient privacy.
This repository contains real system outputs, workflow screenshots, and the full research paper for technical validation.

---

📌 Problem Statement

Most existing systems focus only on cancer detection, not on treatment progress evaluation.
This project fills that gap by:

• Measuring structural similarity between before/after treatment scans
• Detecting positive response, stagnation, or regression risk
• Supporting doctors with objective visual + numerical evidence

---

🧠 Solution Overview

• CNN model classifies cancer severity
• SSIM compares pre-treatment and post-treatment mammograms
• Progress is computed using SSIM score + prediction shift
• A detailed PDF medical report is generated automatically

---

🏗️ System Architecture

📂 Dataset, Model Performance & Database

🧬 Dataset Overview

---

📊 Classification Report

---

📈 Model Accuracy & Loss

---

🗄️ MongoDB Database Cluster

---

⚙️ Tech Stack

• Python
• OpenCV
• TensorFlow / Keras
• SSIM (scikit-image)
• Flask
• MongoDB
• ReportLab (PDF Automation)

---

🖼️ Application Workflow & Screenshots

1️⃣ Home Page

---

2️⃣ User Authentication

---

3️⃣ Mammogram Upload (Pre & Post Treatment)

---

4️⃣ Image Preprocessing

Operations include:

• Grayscale conversion
• Denoising
• Contrast enhancement
• Edge sharpening

---

5️⃣ SSIM Similarity Analysis

Visual heatmap showing structural differences between scans

---

6️⃣ Automated PDF Medical Report

---

📊 Experimental Results

Metric	Value
CNN Accuracy	~92%
SSIM Score Range	0.68 – 0.92
Report Type	Fully Automated PDF

✅ Detects:

• Positive treatment response
• No-response cases
• Possible regression risk

---

📄 Research Paper

📂 Full research paper available here:
👉 https://ijmec.com/index.php/multidisciplinary/article/view/835

---

🚨 Important Disclaimer

• No datasets are uploaded due to medical privacy laws
• No source code is public due to institutional IP policies
• This repository is meant for technical evaluation, interviews, and academic verification

If you need code access for official review, it can be shared under NDA.

---

👤 Author

Syed Sadiq SF
MCA Graduate | AI & Medical Imaging
Skills: Flask, CNN, OpenCV, SSIM, MongoDB, PDF Automation
GitHub: https://github.com/sxdiqsf
LinkedIn: https://linkedin.com/in/sadiq-syed-sf/

---