Rearrange detection.

app.py

@@ -17,6 +17,7 @@ from datetime import datetime
 from tensorflow import image
 from tensorflow.python.keras.models import load_model
 from keras.applications.densenet import DenseNet121
+from keras.layers import Dense, GlobalAveragePooling2D
 from pydicom.pixel_data_handlers.util import apply_voi_lut
 
 # Environment Configuration
@@ -34,13 +35,13 @@ enhancement_type = st.sidebar.selectbox(
     ["Invert", "High Pass Filter", "Unsharp Masking", "Histogram Equalization", "CLAHE"]
 )
 
-H = 224
-W = 224
+H_detection = 224
+W_detection = 224
 
 @st.cache_resource
 def load_model_detection():
-    model = tf.keras.models.load_model("model-detection.h5", compile=False)
-    model.compile(
+    model_detection = tf.keras.models.load_model("files/model-detection.h5", compile=False)
+    model_detection.compile(
         loss={
             "bbox": "mse",
             "class": "sparse_categorical_crossentropy"
@@ -51,17 +52,30 @@ def load_model_detection():
             "class": ['accuracy']
         }
     )
-    return model
+    return model_detection
 
 def preprocess_image(image):
-#     """ Preprocess the image to the required size and normalization. """
-#     W = 320
-#     H = 320
-    image = cv2.resize(image, (W, H))
+    """ Preprocess the image to the required size and normalization. """
+    image = cv2.resize(image, (W_detection, H_detection))
     image = (image - 127.5) / 127.5  # Normalize to [-1, +1]
     image = np.expand_dims(image, axis=0).astype(np.float32)
     return image
 
+def predict(model_detection, image):
+    """ Predict bounding box and label for the input image. """
+    pred_bbox, pred_class = model_detection.predict(image)
+    pred_label_confidence = np.max(pred_class, axis=1)[0]
+    pred_label = np.argmax(pred_class, axis=1)[0]
+    return pred_bbox[0], pred_label, pred_label_confidence
+
+def draw_bbox(image, bbox):
+    """ Draw bounding box on the image. """
+    h, w, _ = image.shape
+    x1, y1, x2, y2 = bbox
+    x1, y1, x2, y2 = int(x1 * w), int(y1 * h), int(x2 * w), int(y2 * h)
+    image = cv2.rectangle(image, (x1, y1), (x2, y2), (255, 0, 0), 2)
+    return image
+
 
 
 # st.title("Chest X-ray Disease Detection")
@@ -369,20 +383,20 @@ def redirect_button(url):
 ###########################################################################################
 
 
-def predict(model_detection, image):
-    """ Predict bounding box and label for the input image. """
-    pred_bbox, pred_class = model_detection.predict(image)
-    pred_label_confidence = np.max(pred_class, axis=1)[0]
-    pred_label = np.argmax(pred_class, axis=1)[0]
-    return pred_bbox[0], pred_label, pred_label_confidence
+# def predict(model_detection, image):
+#     """ Predict bounding box and label for the input image. """
+#     pred_bbox, pred_class = model_detection.predict(image)
+#     pred_label_confidence = np.max(pred_class, axis=1)[0]
+#     pred_label = np.argmax(pred_class, axis=1)[0]
+#     return pred_bbox[0], pred_label, pred_label_confidence
 
-def draw_bbox(image, bbox):
-    """ Draw bounding box on the image. """
-    h, w, _ = image.shape
-    x1, y1, x2, y2 = bbox
-    x1, y1, x2, y2 = int(x1 * w), int(y1 * h), int(x2 * w), int(y2 * h)
-    image = cv2.rectangle(image, (x1, y1), (x2, y2), (255, 0, 0), 2)
-    return image
+# def draw_bbox(image, bbox):
+#     """ Draw bounding box on the image. """
+#     h, w, _ = image.shape
+#     x1, y1, x2, y2 = bbox
+#     x1, y1, x2, y2 = int(x1 * w), int(y1 * h), int(x2 * w), int(y2 * h)
+#     image = cv2.rectangle(image, (x1, y1), (x2, y2), (255, 0, 0), 2)
+#     return image
 
 
 ###########################################################################################
@@ -480,7 +494,7 @@ if uploaded_file is not None:
 
     with col2:
         file_bytes = np.asarray(bytearray(uploaded_file.read()), dtype=np.uint8)
-        image = cv2.imdecode(file_bytes, 1) 
+        image = cv2.imdecode(file_bytes, 1)
         if st.button('Auto Detect'):
             st.write("Processing...")
             input_image = preprocess_image(image)