App change

README.md

@@ -10,5 +10,36 @@ pinned: false
 license: mit
 short_description: 🩺 Bone Age Prediction Gradio App
 ---
+# 🩺 Bone Age Prediction Gradio App
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+This Hugging Face Space provides an interactive demo for bone age estimation from hand X-ray images, based on the [bone-age-resnet-80m](https://huggingface.co/Adilbai/bone-age-resnet-80m) deep learning model (ResNet152, finetuned on the RSNA Pediatric Bone Age dataset).
+
+## 🚀 What does this app do?
+
+- **Upload a hand X-ray** (PNG/JPG).
+- **Select the patient's gender** (Male/Female).
+- **Get an instant prediction** of bone age in months (and years/months format) using a state-of-the-art neural network.
+
+## 🧠 Model Details
+
+- **Architecture:** ResNet152 + custom head (≈80M parameters)
+- **Input:** 256x256 hand X-ray image & gender
+- **Output:** Bone age (months)
+- **Training Data:** [RSNA Bone Age Challenge](https://www.kaggle.com/datasets/kmader/rsna-bone-age)
+- **Model Card:** [bone-age-resnet-80m](https://huggingface.co/Adilbai/bone-age-resnet-80m)
+
+## 🌟 How to use
+
+1. Upload a clear hand X-ray image (preferably as PNG).
+2. Select the appropriate gender.
+3. Press "Submit" to get the predicted bone age.
+
+> **Note:** This app is for educational and research purposes only. Not for clinical use.
+
+## 🏷️ Citation
+
+If you use this demo or model, please cite the [RSNA Bone Age dataset](https://www.kaggle.com/datasets/kmader/rsna-bone-age) and this Hugging Face Space/model.
+
+---
+
+Built with ❤️ using Gradio and Hugging Face Spaces.

app.py

@@ -0,0 +1,86 @@
+import gradio as gr
+from PIL import Image
+import torch
+import numpy as np
+import requests
+from io import BytesIO
+from torchvision import transforms
+import onnxruntime as ort
+
+# ======================
+# Model & Preprocessing
+# ======================
+MODEL_ONNX_URL = "https://huggingface.co/Adilbai/bone-age-resnet-80m/resolve/main/resnet_bone_age_80m.onnx"
+
+def download_model(url, filename):
+    if not os.path.exists(filename):
+        print(f"Downloading model from {url}")
+        r = requests.get(url)
+        with open(filename, "wb") as f:
+            f.write(r.content)
+
+import os
+MODEL_PATH = "resnet_bone_age_80m.onnx"
+download_model(MODEL_ONNX_URL, MODEL_PATH)
+
+# Set up ONNX session
+ort_session = ort.InferenceSession(MODEL_PATH)
+
+# Define image preprocessing (must match training)
+transform = transforms.Compose([
+    transforms.Resize((256, 256)),
+    transforms.ToTensor(),
+    transforms.Normalize([0.5]*3, [0.5]*3)
+])
+
+# ======================
+# Inference Function
+# ======================
+def predict_bone_age(image, gender):
+    """
+    image: PIL.Image
+    gender: string ("Male" or "Female")
+    """
+    # Preprocess image
+    img_tensor = transform(image).unsqueeze(0).numpy()
+    # Gender: 0=male, 1=female
+    gender_val = 0.0 if gender.lower() == "male" else 1.0
+    gender_tensor = np.array([[gender_val]], dtype=np.float32)
+
+    # ONNX inference
+    outputs = ort_session.run(None, {"image": img_tensor, "gender": gender_tensor})
+    pred_age = outputs[0][0]
+
+    # Display as years and months
+    years = int(pred_age // 12)
+    months = int(pred_age % 12)
+    result_str = (
+        f"Predicted Bone Age: **{pred_age:.1f} months**  \n"
+        f"≈ {years} years, {months} months"
+    )
+    return result_str
+
+# ======================
+# Gradio UI
+# ======================
+app_title = "Bone Age Prediction from Hand X-ray"
+app_desc = """
+Upload a hand X-ray image and select the patient's gender. This app will predict the bone age (in months) using a powerful deep learning model.
+- Model: [bone-age-resnet-80m](https://huggingface.co/Adilbai/bone-age-resnet-80m)
+- Data: RSNA Pediatric Bone Age Challenge
+- **For research/educational use only.**
+"""
+iface = gr.Interface(
+    fn=predict_bone_age,
+    inputs=[
+        gr.Image(type="pil", label="Hand X-ray Image"),
+        gr.Radio(["Male", "Female"], label="Gender")
+    ],
+    outputs=gr.Markdown(label="Prediction"),
+    title=app_title,
+    description=app_desc,
+    allow_flagging="never"
+)
+
+if __name__ == "__main__":
+    iface.launch()

requirements.txt

@@ -0,0 +1,6 @@
+gradio
+torch
+torchvision
+onnxruntime
+Pillow
+requests