ResNet-50 Model for Acne Severity Classification

This is a fine-tuned ResNet-50 model designed to classify the severity of acne from medical images into five categories (Severity 1 to Severity 5). The model leverages transfer learning on ResNet-50 pre-trained on ImageNet and adapts it for acne severity classification tasks.

Model Overview

Model Details

Key Features

Input: Medical images of acne-affected skin.
Output: Severity classification with one of the following labels:
- level0 (No acne or minimal severity)
- level1 (Mild severity)
- level2 (Moderate severity)
- level3 (Severe or advanced acne)

Training Details

Framework: PyTorch
Base Model: ResNet-50 (pretrained on ImageNet)
Dataset: A balanced dataset of acne images annotated with severity levels (Severity 1 to 5).
Preprocessing: Images resized to 224x224 pixels, normalized using ImageNet statistics (mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225]).
Optimizer: Adam with a learning rate of 0.001.
Loss Function: CrossEntropyLoss.
Epochs: 10.
Validation Accuracy: 0.85 (on a held-out validation set).

How to Use the Model

Intended Use

This model is intended for educational purposes and demonstrates image classification for medical images. It should not be used for clinical decision-making without further validation.

Example Usage

You can use this model via the Hugging Face Transformers pipeline for inference. Ensure you have the transformers library installed:

pip install transformers

Hugging Face Inference API

You can use the model via the Hugging Face Inference API by sending an image encoded in base64. Here’s an example:

curl -X POST https://api-inference.huggingface.co/models/YOUR_MODEL_NAME \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"inputs": "BASE64_ENCODED_IMAGE"}'