███████╗██╗ ██╗██╗███╗ ██╗ ██████╗ ███████╗████████╗███████╗ ██████╗████████╗
██╔════╝██║ ██╔╝██║████╗ ██║ ██╔══██╗██╔════╝╚══██╔══╝██╔════╝██╔════╝╚══██╔══╝
███████╗█████╔╝ ██║██╔██╗ ██║ ██║ ██║█████╗ ██║ █████╗ ██║ ██║
╚════██║██╔═██╗ ██║██║╚██╗██║ ██║ ██║██╔══╝ ██║ ██╔══╝ ██║ ██║
███████║██║ ██╗██║██║ ╚████║ ██████╔╝███████╗ ██║ ███████╗╚██████╗ ██║
╚══════╝╚═╝ ╚═╝╚═╝╚═╝ ╚═══╝ ╚═════╝ ╚══════╝ ╚═╝ ╚══════╝ ╚═════╝ ╚═╝
A CNN-powered clinical decision support tool for early-stage skin cancer diagnosis
"Early detection saves lives. This system brings AI-powered dermatology to anyone with a camera."
Analysis Dashboard
Prediction Result
Model Performance
Skin cancer is among the most diagnosed cancers worldwide - melanoma alone accounts for the majority of skin cancer deaths. The five-year survival rate drops from ~99% when detected early to just ~27% when detected late.
Yet access to dermatologists remains severely limited in rural and low-income regions globally. This project bridges that gap:
- No clinical equipment required - works with any camera-captured image
- Instant AI-assisted diagnosis - results in seconds through a web browser
- Research-grade model - trained on the ISIC Archive, peer-reviewed and conference-presented
| Feature | Details |
|---|---|
| 95.4% Accuracy | CNN classifier with Softmax output trained on ISIC archive data |
| Spectral Analysis | Fourier transform integrated into the feature extraction pipeline |
| 3 Cancer Types Detected | Melanoma · Basal Cell Carcinoma · Squamous Cell Carcinoma |
| Browser-Based UI | Flask-powered interface - upload image, get instant prediction |
| Secure Authentication | Argon2 password hashing + session management |
| Full EDA Pipeline | Preprocessing, lesion segmentation, and augmentation |
Input Image (RGB)
│
▼
┌──────────────────┐
│ Preprocessing │ ← Resize · Normalize · Augment
└────────┬─────────┘
│
▼
┌──────────────────────────┐
│ Fourier Spectral Layer │ ← Frequency-domain feature extraction
└────────┬─────────────────┘
│
▼
┌──────────────────┐
│ CNN Backbone │ ← Conv2D → MaxPool → BatchNorm (stacked)
└────────┬─────────┘
│
▼
┌──────────────────┐
│ Dense Layers │ ← Dropout regularization
└────────┬─────────┘
│
▼
┌──────────────────┐
│ Softmax Output │ → Melanoma / BCC / SCC
└──────────────────┘
Design highlights:
- Fourier spectral layer captures texture patterns invisible to standard CNNs
- Batch normalization after each convolutional block stabilizes training
- Dropout layers prevent overfitting on medical imaging data
┌──────────────┬─────────────────────────────────────────┐
│ Layer │ Technologies │
├──────────────┼─────────────────────────────────────────┤
│ Frontend │ HTML5, CSS3, Jinja2 Templates │
│ Backend │ Python 3.8+, Flask │
│ ML Model │ TensorFlow / Keras (CNN) │
│ Database │ SQLite │
│ Libraries │ NumPy, OpenCV, Matplotlib, joblib │
│ Security │ Argon2 (password hashing) │
└──────────────┴─────────────────────────────────────────┘
smart-skin-care/
│
├── model/
│ └── model.h5 # Trained CNN weights
│
├── static/
│ └── uploads/ # User-uploaded images (sandboxed)
│
├── templates/
│ ├── index.html # Landing page
│ ├── about.html # Project info
│ ├── register.html # User registration
│ ├── login.html # User login
│ ├── predict.html # Image upload & prediction
│ └── result.html # Diagnosis result display
│
├── app.py # Flask application entry point
├── create_database.py # SQLite schema setup
├── utils.py # Preprocessing & helper functions
└── requirements.txt
- Python 3.8+
- pip / virtualenv
# 1. Clone the repository
git clone https://github.com/your-username/smart-skin-care.git
cd smart-skin-care
# 2. Create and activate virtual environment
python -m venv venv
source venv/bin/activate # macOS/Linux
# venv\Scripts\activate # Windows
# 3. Install dependencies
pip install -r requirements.txt
# 4. Initialize the database
python create_database.py
# 5. Launch the application
python app.pyOpen http://localhost:5000 in your browser.
┌──────────────────────────────────────────────────┐
│ Classification Report │
├──────────────────────────┬───────────────────────┤
│ Metric │ Score │
├──────────────────────────┼───────────────────────┤
│ Overall Accuracy │ 95.4% │
│ Dataset │ ISIC Archive + │
│ │ Dermatology DB │
│ Classes │ 3 │
│ Classifier │ CNN + Softmax │
└──────────────────────────┴───────────────────────┘
Melanoma ████████████████████ 95%+
BCC ████████████████████ 95%+
SCC ████████████████████ 95%+
The model was trained and validated on the ISIC Archive, the largest publicly available dataset for skin lesion analysis.
| Measure | Implementation |
|---|---|
| Password Security | Argon2 hashing - memory-hard, phishing-resistant |
| Session Management | Server-side validation, no client-stored secrets |
| File Isolation | Uploaded images sandboxed in /static/uploads/ |
| Input Validation | File type and size checks before model inference |
| Field | Details |
|---|---|
| Journal | International Journal of Scientific Research in Engineering and Management (IJSREM) |
| Volume / Issue | Vol. 09, Issue 11 - November 2025 |
| SJIF Impact Factor | 8.586 |
| ISSN | 2582-3930 |
| DOI | 10.55041/IJSREM54416 |
| Authors | Dr. G. Kavitha · S. Harippriya · R. Lakshana |
Download Published Paper · View on IJSREM · Drive - Paper & References
Dr. G. Kavitha · S. Harippriya · R. Lakshana
"Smart Skin Care: Deep Learning In Skin Cancer Detection"
National Conference on Innovative Trends in Technologies (NCITT'25)
Organized by Computer Society of India - KEC Student Branch
Perundurai, Erode, Tamil Nadu, India · 01 March 2025
| Name | Roll No. | Role |
|---|---|---|
| S. Harippriya | 21CS089 | Developer & Researcher |
| R. Lakshana | 21CS106 | Developer & Researcher |
Guide: Dr. G. Kavitha, M.S (By Research), Ph.D
Department of Computer Science and Engineering
Muthayammal Engineering College
Potential directions for future work:
- Expand to 7-class HAM10000 classification
- GRAD-CAM heatmap overlay for explainability
- Mobile-responsive PWA for field deployment
- REST API endpoint for third-party EMR integration
- Model quantization for edge device inference
Click to expand full IEEE citation list
[1] Azhar Imran, Arslan Nasir, B. Mohammed, D. Foong, and A. Abbosh, "Benign and malignant skin lesions: Dielectric characterization modelling and analysis in frequency band 1 to 14 GHz," IEEE Trans., vol. 70, no. 2, pp. 628–639, Feb. 2022.
[2] C. P. Davis, Rosacea, Acne, Shingles, Covid-19 Rashes: Common Adult Skin Diseases, 2020.
[3] Dipu Chandra Malo et al., "Skin Cancer Detection using Convolutional Neural Network," 2022 IEEE 12th Annual Computing and Communication Workshop and Conference (CCWC), pp. 0169–0176, 2022.
[4] Krishna Mridha and Md. Mezbah Uddin, "A Data-Driven Predictive Model for Speed Control in Automotive Safety Applications," IEEE Sensors Journal, vol. 22, no. 23, pp. 23258–23266, Dec. 2023.
[5] Lubna Riaz and Anca D. Jurcut, "FeduLPM: Federated Unsupervised Learning-Based Predictive Model for Speed Control in Customizable Automotive Variants," IEEE Sensors Journal, vol. 23, no. 13, pp. 14700–14708, July 2023.
[6] M. Chen et al., "AI-skin: Skin disease recognition based on self-learning and wide data collection through a closed-loop framework," Inf. Fusion, vol. 54, pp. 1–9, Feb. 2020.
[7] M. Goyal, T. Knackstedt, S. Yan, and S. Hassanpour, "Artificial intelligence-based image classification methods for diagnosis of skin cancer: Challenges and opportunities," Comput. Biol. Med., vol. 127, Dec. 2020.
[8] Raissa Schiavoni and Gennaro Maietta, "In vivo human skin dielectric properties characterization and statistical analysis at frequencies from 1 to 30 GHz," IEEE Trans. Instrum. Meas., vol. 70, 2023.
[9] S. A. R. Naqvi et al., "Benign and malignant skin lesions: Dielectric characterization modelling and analysis in frequency band 1 to 14 GHz," IEEE Trans. Biomed. Eng., vol. 70, no. 2, pp. 628–639, Feb. 2023.
[10] S. A. R. Naqvi et al., "In vivo human skin dielectric properties characterization and statistical analysis at frequencies from 1 to 30 GHz," IEEE Trans. Instrum. Meas., vol. 70, 2021.
[11] S. Samsudeen and G. Senthil Kumar, "A Data-Driven Predictive Model for Speed Control in Automotive Safety Applications," IEEE Sensors Journal, vol. 22, no. 23, pp. 23258–23266, Dec. 2022.
[12] S. Samsudeen and G. S. Kumar, "FeduLPM: Federated Unsupervised Learning-Based Predictive Model for Speed Control in Customizable Automotive Variants," IEEE Sensors Journal, vol. 23, no. 13, pp. 14700–14708, July 2023.
[13] S. S. Chaturvedi, J. V. Tembhurne, and T. Diwan, "A multi-class skin cancer classification using deep convolutional neural networks," Multimedia Tools Appl., vol. 79, nos. 39–40, pp. 28477–28498, Oct. 2020.
[14] Yessi Jusman et al., "Performance of multi layer perceptron and deep neural networks in skin cancer classification," 2021 IEEE 3rd Global Conference on Life Sciences and Technologies (LifeTech), pp. 534–538, 2021.
This tool is intended as a clinical decision support aid only and does not replace professional medical diagnosis. Always consult a qualified dermatologist for medical advice.
Muthayammal Engineering College · 2025