| | --- |
| | license: apache-2.0 |
| | datasets: |
| | - Nagabu/HAM10000 |
| | metrics: |
| | - accuracy |
| | pipeline_tag: image-classification |
| | library_name: transformers |
| | --- |
| | |
| | # Dermacare Skin Lesion Classification |
| |
|
| | Dermacare is a skin lesion classification model built using Keras. It classifies dermatoscopic images into various types of skin lesions, aiding in the early detection of skin cancer. |
| |
|
| | ## Model Architecture |
| | The model is a Convolutional Neural Network (CNN) trained on the [HAM10000 dataset](https://www.kaggle.com/datasets/ultralytics/ham10000) with the following key specifications: |
| |
|
| | - **Input**: 224x224 RGB images |
| | - **Architecture**: Keras-based CNN |
| | - **Output**: 7-class classification for different types of skin lesions |
| |
|
| | ## Usage Example |
| | To use the model for predictions, send an image to the inference endpoint as shown below: |
| |
|
| | ```python |
| | import requests |
| | |
| | API_URL = "https://api-inference.huggingface.co/models/sreejith782/Dermacare_Skin_Lesion_classification" |
| | headers = {"Authorization": "Bearer YOUR_HUGGING_FACE_TOKEN"} |
| | |
| | response = requests.post(API_URL, headers=headers, files={"inputs": open("path_to_image.jpg", "rb")}) |
| | print(response.json()) |
