PressCare - Pressure Ulcer Classification Model

This model classifies pressure ulcer wounds into 7 categories:

• Invalid
• SDTI (Suspected Deep Tissue Injury)
• Stage I
• Stage II
• Stage III
• Stage IV
• Unstageable

Model Details

• Architecture: RexNet-150
• Input size: 224x224
• Framework: PyTorch with timm

Usage with Inference Endpoints

The model includes a custom handler for easy deployment on Hugging Face Inference Endpoints.

API Usage

import requests
import base64

API_URL = "https://YOUR-ENDPOINT.endpoints.huggingface.cloud"
headers = {"Authorization": "Bearer YOUR_TOKEN"}

with open("wound_image.jpg", "rb") as f:
    image_base64 = base64.b64encode(f.read()).decode()

response = requests.post(
    API_URL,
    headers=headers,
    json={"inputs": {"image": image_base64}}
)

print(response.json())

Response Format

{
  "predictions": [
    {"label": "Stage_III", "score": 0.85},
    {"label": "Stage_IV", "score": 0.10},
    {"label": "Stage_II", "score": 0.03}
  ],
  "probabilities": {
    "Invalid": 0.01,
    "SDTI": 0.005,
    "Stage_I": 0.003,
    "Stage_II": 0.03,
    "Stage_III": 0.85,
    "Stage_IV": 0.10,
    "Unstageable": 0.002
  },
  "predicted_class": "Stage_III",
  "confidence": 0.85
}

Disclaimer

This model is for educational and research purposes only. Always consult healthcare professionals for medical decisions. EOF < /dev/null