Skin Cancer Image Classification Model
Introduction
This model is designed for the classification of skin cancer images into various categories including benign keratosis-like lesions, basal cell carcinoma, actinic keratoses, vascular lesions, melanocytic nevi, melanoma, and dermatofibroma.
Model Overview
- Model Architecture: Vision Transformer (ViT)
- Pre-trained Model: Google's ViT with 16x16 patch size and trained on ImageNet21k dataset
- Modified Classification Head: The classification head has been replaced to adapt the model to the skin cancer classification task.
Dataset
- Dataset Name: Skin Cancer Dataset
- Source: Marmal88's Skin Cancer Dataset on Hugging Face
- Classes: Benign keratosis-like lesions, Basal cell carcinoma, Actinic keratoses, Vascular lesions, Melanocytic nevi, Melanoma, Dermatofibroma
Training
- Optimizer: Adam optimizer with a learning rate of 1e-4
- Loss Function: Cross-Entropy Loss
- Batch Size: 32
- Number of Epochs: 5
Evaluation Metrics
- Train Loss: Average loss over the training dataset
- Train Accuracy: Accuracy over the training dataset
- Validation Loss: Average loss over the validation dataset
- Validation Accuracy: Accuracy over the validation dataset
Results
- Epoch 5/5, Train Loss: 0.1208, Train Accuracy: 0.9614, Val Loss: 0.1000, Val Accuracy: 0.9695
Conclusion
The model demonstrates good performance in classifying skin cancer images into various categories. Further fine-tuning or experimentation may improve performance on this task.
