Initial Commit

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+model/variables/variables.data-00000-of-00001 filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,79 @@
+---
+language: en
+license: mit
+tags:
+  - tensorflow
+  - image-classification
+  - medical-imaging
+  - vascular-dementia
+datasets:
+  - custom
+---
+
+# EfficientNet Model for Vascular Dementia Detection
+
+This model was trained to detect Vascular Dementia (VAD) from MRI scans. It uses an EfficientNet architecture fine-tuned on a custom dataset of brain MRI images.
+
+## Model Description
+
+- **Model Type:** EfficientNet
+- **Task:** Binary classification (VAD-Demented vs. Non-Demented)
+- **Input:** MRI brain scans (224x224 RGB)
+- **Output:** Binary classification with confidence score
+
+## Usage
+
+```python
+import tensorflow as tf
+import numpy as np
+from PIL import Image
+
+# Load the model
+model = tf.keras.models.load_model("path/to/downloaded/model")
+
+# Preprocess your image
+image = Image.open("path/to/your/mri.jpg")
+image = image.resize((224, 224))
+image_array = np.array(image) / 255.0
+image_array = np.expand_dims(image_array, axis=0)
+
+# Get prediction
+prediction = model.predict(image_array)
+predicted_class = "VAD-Demented" if prediction[0][0] > 0.5 else "Non-Demented"
+confidence = prediction[0][0] * 100 if prediction[0][0] > 0.5 else (1 - prediction[0][0]) * 100
+
+print(f"Prediction: {predicted_class}")
+print(f"Confidence: {confidence:.2f}%")
+```
+
+## API Usage
+
+This model can be used directly with the Hugging Face Inference API:
+
+```python
+import requests
+import base64
+from PIL import Image
+import io
+
+# Convert image to base64
+image = Image.open("path/to/your/mri.jpg")
+buffered = io.BytesIO()
+image.save(buffered, format="JPEG")
+img_str = base64.b64encode(buffered.getvalue()).decode()
+
+# API endpoint
+API_URL = "https://api-inference.huggingface.co/models/thakshana02/vad-efficientnet-model"
+
+# API headers with your token
+headers = {"Authorization": "Bearer YOUR_TOKEN"}
+
+# Make prediction request
+response = requests.post(API_URL, headers=headers, json={"inputs": {"image": img_str}})
+result = response.json()
+print(result)
+```
+
+## Limitations
+
+This model is intended for research purposes only and should not be used for clinical diagnosis without proper validation by healthcare professionals.

app.py

@@ -0,0 +1,65 @@
+import tensorflow as tf
+import numpy as np
+from PIL import Image
+import io
+import base64
+import json
+
+class MRIClassifier:
+    def __init__(self):
+        # Load the model
+        self.model = tf.keras.models.load_model("./model")
+        self.img_size = (224, 224)  # EfficientNet standard size
+        print(f"Model loaded with input size: {self.img_size}")
+        
+    def preprocess(self, image):
+        # Resize image to expected input shape
+        image = image.resize(self.img_size)
+        
+        # Convert to RGB if not already
+        if image.mode != "RGB":
+            image = image.convert("RGB")
+            
+        # Convert to array and normalize
+        image_array = np.array(image) / 255.0
+        
+        # Add batch dimension
+        image_array = np.expand_dims(image_array, axis=0)
+        
+        return image_array
+        
+    def predict(self, image_bytes):
+        # Convert bytes to image
+        image = Image.open(io.BytesIO(image_bytes))
+        
+        # Preprocess image
+        image_array = self.preprocess(image)
+        
+        # Make prediction
+        prediction = self.model.predict(image_array)
+        pred_value = float(prediction[0][0])
+        
+        # Format result
+        predicted_class = "VAD-Demented" if pred_value > 0.5 else "Non-Demented"
+        
+        return [
+            {
+                "label": predicted_class,
+                "score": pred_value
+            }
+        ]
+
+# Initialize classifier
+classifier = MRIClassifier()
+
+def inference(model_inputs):
+    # Handle both direct image inputs and base64 encoded images
+    if isinstance(model_inputs, dict) and "image" in model_inputs:
+        # Base64 encoded image
+        image_bytes = base64.b64decode(model_inputs["image"])
+    else:
+        # Direct image bytes (from API)
+        image_bytes = model_inputs
+        
+    results = classifier.predict(image_bytes)
+    return results

best_model_efficientnet_vad2.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ff88252154c2d64ddfc42bf4bfa20a142ec8f2f26a385dfd1afd4c0fea0944d
+size 36286472

model.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f789cb5ad86c4b708307dfd270a317c8586d186b1eab05e04abc9d934a76ee31
+size 17522648

model/fingerprint.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:609519a3390dd5c72fcf5babaea7d589755aea9d77ac6c3151334f51dccc0c19
+size 55

model/keras_metadata.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f027ee845a0d0b9be32a779ccedc9b86b11bde0b48c73991b2de3e7e6879280
+size 524632

model/saved_model.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b3ca74fb08fbffe0b615e47393732cc2928b5f835127c84ad7f72cff2f7d671d
+size 3865326

model/variables/variables.data-00000-of-00001

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6c9676b6b5de2cba69357b13569f77d88eede30da8e161ef6e95c8f44bb2e74
+size 17045849

requirements.txt

@@ -0,0 +1,3 @@
+tensorflow>=2.4.0
+pillow
+numpy