|
|
--- |
|
|
license: mit |
|
|
tags: |
|
|
- medical |
|
|
- cnn |
|
|
- dermatology |
|
|
- melanoma |
|
|
--- |
|
|
|
|
|
# simpleNet: Skin Lesion Classification (Benign vs Malignant) |
|
|
|
|
|
**simpleNet** is a lightweight Convolutional Neural Network (CNN) model for binary skin lesion classification, distinguishing between **benign** and **malignant (melanoma)** cases. |
|
|
|
|
|
## Model Description |
|
|
- Architecture: Custom CNN with 3 convolutional blocks and regularization (Batch Normalization, Dropout, L2). |
|
|
- Input size: **224 × 224 × 3 RGB images**. |
|
|
- Output: 2 classes → `["benign", "malignant"]`. |
|
|
- Trained on a combination of curated skin lesion datasets, including **ISIC** samples. |
|
|
|
|
|
The model is optimized for **generalization** and has been validated with external images, showing robust performance for melanoma detection. |
|
|
|
|
|
## Intended Use |
|
|
- **Educational and research purposes**. |
|
|
- Demonstrating the potential of CNNs for medical imaging. |
|
|
- Not intended for clinical use without further validation and regulatory approval. |
|
|
|
|
|
|
