u1

Dockerfile

@@ -0,0 +1,15 @@
+FROM python:3.9
+
+COPY requirements.txt .
+
+RUN pip install -r requirements.txt
+
+RUN apt-get update && apt-get install -y libgl1-mesa-glx
+
+RUN python -c "import cv2; import numpy"
+
+COPY . .
+
+EXPOSE 7860
+
+CMD ["flask", "run", "--host=0.0.0.0", "--port=7860"]

app.py

@@ -0,0 +1,49 @@
+from flask import Flask, render_template, request
+from model import btd
+import numpy as np
+import cv2
+
+app = Flask(__name__)
+
+@app.route("/", methods=["GET", "POST"])
+def index():
+    message = "Welcome to the API! Here are the available endpoints:"
+    fm_url = request.host_url + "fusion-predicition"
+    response = {
+        "message": message,
+        "endpoints": {
+            "fused_prediction": fm_url,
+        }
+    }
+    return response, 200
+
+@app.route("/fusion-predicition", methods=["GET", "POST"])
+def test():
+    result = None
+    if request.method == "POST":
+        if 'img1' not in request.files or 'img2' not in request.files or 'img3' not in request.files:
+            return "No file uploaded for Tumor prediction", 400
+        
+        img1 = request.files['img1']
+        img2 = request.files['img2']
+        img3 = request.files['img3']
+
+        img_array1 = cv2.imdecode(np.frombuffer(
+            img1.read(), np.uint8), cv2.IMREAD_COLOR)
+        img_array2 = cv2.imdecode(np.frombuffer(
+            img2.read(), np.uint8), cv2.IMREAD_COLOR)
+        img_array3 = cv2.imdecode(np.frombuffer(
+            img3.read(), np.uint8), cv2.IMREAD_COLOR)
+        model = btd.FusedFuctionModel(img_array1, img_array2, img_array3)
+
+        response = {
+            "Tumor Predicition": model
+        }
+
+        result = response
+        
+    return result
+
+
+if __name__ == "__main__":
+    app.run(debug=True)

model/btd.py

@@ -0,0 +1,97 @@
+import cv2
+import numpy as np
+from keras.models import Sequential , Model
+from tensorflow.keras.applications.vgg19 import VGG19
+from tensorflow.keras.applications.mobilenet import MobileNet
+from keras.layers import Dense
+from tensorflow.keras.applications.inception_v3 import InceptionV3
+from keras.layers import  Concatenate, Flatten
+
+
+def FusedFuctionModel(img1, img2, img3):
+    inception_model_1 = Sequential()
+
+    inception = InceptionV3(weights="imagenet",
+                            include_top=False,
+                            input_shape=(128, 128, 3))
+
+    for layer1 in inception.layers:
+        layer1.trainable = False
+
+    inception_model_1.add(inception)
+
+    inception_model_1.add(Flatten())
+    inception_model_1.add(Dense(1024, activation="relu"))
+    inception_model_1.add(Dense(1024, activation="relu"))
+    inception_model_1.add(Dense(512, activation="relu"))
+    inception_model_1.add(Dense(128, activation="relu"))
+
+    mobilenet_model_2 = Sequential()
+
+    mobilenet = MobileNet(weights="imagenet",
+                          include_top=False,
+                          input_shape=(128, 128, 3))
+
+    for layer2 in mobilenet.layers:
+        layer2.trainable = False
+
+    mobilenet_model_2.add(mobilenet)
+
+    mobilenet_model_2.add(Flatten())
+    mobilenet_model_2.add(Dense(1024, activation="relu"))
+    mobilenet_model_2.add(Dense(1024, activation="relu"))
+    mobilenet_model_2.add(Dense(512, activation="relu"))
+    mobilenet_model_2.add(Dense(128, activation="relu"))
+
+    vgg19_model_3 = Sequential()
+
+    vgg19 = VGG19(weights="imagenet",
+                  include_top=False,
+                  input_shape=(128, 128, 3))
+
+    for layer3 in vgg19.layers:
+        layer3.trainable = False
+
+    vgg19_model_3.add(vgg19)
+
+    vgg19_model_3.add(Flatten())
+    vgg19_model_3.add(Dense(1024, activation="relu"))
+    vgg19_model_3.add(Dense(1024, activation="relu"))
+    vgg19_model_3.add(Dense(512, activation="relu"))
+    vgg19_model_3.add(Dense(128, activation="relu"))
+
+    merged = Concatenate()([inception_model_1.output, mobilenet_model_2.output, vgg19_model_3.output])
+
+    output = Dense(3, activation='softmax')(merged)
+
+    model_final = Model(inputs=[inception_model_1.input, mobilenet_model_2.input, vgg19_model_3.input],
+                        outputs=[output])
+
+    model_final.compile(loss='sparse_categorical_crossentropy',
+                        optimizer='adam',
+                        metrics=["accuracy"])
+
+    model_final.load_weights("weights/fused-model-weights.h5")
+
+    img1 = cv2.resize(img1, (128, 128))
+    img2 = cv2.resize(img2, (128, 128))
+    img3 = cv2.resize(img3, (128, 128))
+
+    image1 = np.expand_dims(img1, axis=0)
+    image2 = np.expand_dims(img2, axis=0)
+    image3 = np.expand_dims(img3, axis=0)
+
+    predictions = model_final.predict([image1, image2, image3])
+
+    predicted_classes = predictions.argmax(axis=-1)
+
+    result = str(predicted_classes)
+
+    if result == '[0]':
+        final_result = 'glioma'
+    elif result == '[1]':
+        final_result = 'meningimoa'
+    elif result == '[2]':
+        final_result = 'pituitary'
+
+    return final_result

weights/fused-model-weights.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac2b34465910c234805537a09b89637c575e263d013a3404327e4eec26037f9c
+size 334668168