Update app.py

app.py

@@ -29,7 +29,7 @@ bucket_load = storage_client.bucket(bucket_name_load)
 model_path = os.path.join("model.h5")
 model = tf.keras.models.load_model(model_path)
 
-H, W = 512, 512
+H, W = 320, 320
 
 test_samples_folder = 'object_detection_test_samples'
 
@@ -51,7 +51,7 @@ df = pd.read_excel('BBox_List_2017.xlsx')
 labels_dict = dict(zip(df['Image Index'], df['Finding Label']))
 
 def predict(image):
-    H, W = 512, 512
+    H, W = 320, 320
 
     image_resized = cv2.resize(image, (W, H))
     image_normalized = (image_resized - 127.5) / 127.5
@@ -181,7 +181,7 @@ def upload_folder_images(original_image_path, enhanced_image_path):
     upload_to_gcs(enhanced_dicom_bytes, folder_name + '/' + enhancement_name + '.dcm', content_type='application/dicom')
 
 
-def get_mean_std_per_batch(image_path, df, H=512, W=512):
+def get_mean_std_per_batch(image_path, df, H=320, W=320):
     sample_data = []
     for idx, img in enumerate(df.sample(100)["Image Index"].values):
         # path = image_dir + img
@@ -192,7 +192,7 @@ def get_mean_std_per_batch(image_path, df, H=512, W=512):
     std = np.std(sample_data[0])
     return mean, std
 
-def load_image(img_path, preprocess=True, height=512, width=512):
+def load_image(img_path, preprocess=True, height=320, width=320):
     mean, std = get_mean_std_per_batch(img_path, df, height, width)
     x = keras.utils.load_img(img_path, target_size=(height, width))
     x = keras.utils.img_to_array(x)
@@ -225,7 +225,7 @@ def grad_cam(input_model, img_array, cls, layer_name):
     for index, w in enumerate(weights):
         cam += w * output[:, :, index]
 
-    cam = cv2.resize(cam.numpy(), (512, 512), cv2.INTER_LINEAR)
+    cam = cv2.resize(cam.numpy(), (320, 320), cv2.INTER_LINEAR)
     cam = np.maximum(cam, 0)
     cam = cam / cam.max()