diff --git a/src/dopt-sensor-anomalies/__init__.py b/src/dopt_sensor_anomalies/__init__.py
similarity index 100%
rename from src/dopt-sensor-anomalies/__init__.py
rename to src/dopt_sensor_anomalies/__init__.py
diff --git a/src/dopt_sensor_anomalies/constants.py b/src/dopt_sensor_anomalies/constants.py
new file mode 100644
index 0000000..fe344eb
--- /dev/null
+++ b/src/dopt_sensor_anomalies/constants.py
@@ -0,0 +1,10 @@
+from typing import Final
+
+THRESHOLD_BW: Final[int] = 63  # threshold to distringuish black (electrodes) and white areas
+# model_path = [
+#     r"C:\Users\demon\Documents\EKF\Modelle\patchcore_model_links.pth",
+#     r"C:\Users\demon\Documents\EKF\Modelle\patchcore_model_rechts.pth",
+# ]  # path to anomaly detection models
+BACKBONE: Final[str] = "resnet18"  # parameters for AI model
+LAYERS: Final[tuple[str, str]] = ("layer1", "layer2")
+RATIO: Final[float] = 0.05
diff --git a/src/dopt-sensor-anomalies/detection.py b/src/dopt_sensor_anomalies/detection.py
similarity index 97%
rename from src/dopt-sensor-anomalies/detection.py
rename to src/dopt_sensor_anomalies/detection.py
index 51a3e61..3f4ded9 100644
--- a/src/dopt-sensor-anomalies/detection.py
+++ b/src/dopt_sensor_anomalies/detection.py
@@ -19,21 +19,20 @@ from scipy.spatial import distance as dist
 from torch import as_tensor, cuda, device, float32, load, no_grad
 from torchvision.transforms.v2.functional import to_dtype, to_image
 
-# input parameters
+from dopt_sensor_anomalies import constants as const
+
+# input parameters: user-defined
 file_path = r"C:\Users\demon\Documents\EKF\Analyse_fuer_Florian\bild2.bmp"
 pixelsPerMetricX = 0.251
 pixelsPerMetricY = 0.251
 
 
-# internal parameters
-schwellwert = 63  # threshold to distringuish black (electrodes) and white areas
+# internal parameters - configuration
+schwellwert = 63  # threshold to distinguish black (electrodes) and white areas
 model_path = [
     r"C:\Users\demon\Documents\EKF\Modelle\patchcore_model_links.pth",
     r"C:\Users\demon\Documents\EKF\Modelle\patchcore_model_rechts.pth",
 ]  # path to anomaly detection models
-backbone = "resnet18"  # parameters for AI model
-layers = ["layer1", "layer2"]
-ratio = 0.05
 
 
 # measuring
@@ -114,7 +113,7 @@ def measure_length(file_path, pixelsPerMetricX, pixelsPerMetricY):
 
     # change colours in the image to black and white
     gray = cv2.cvtColor(cropped, cv2.COLOR_BGR2GRAY)
-    _, binary = cv2.threshold(gray, schwellwert, 255, cv2.THRESH_BINARY)
+    _, binary = cv2.threshold(gray, const.THRESHOLD_BW, 255, cv2.THRESH_BINARY)
 
     # perform edge detection, identify rectangular shapes
     kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5))
@@ -350,7 +349,9 @@ def anomaly_detection(file_path, data_csv, sensor_images):
     folder_path = path.dirname(file_path)
 
     # reconstruct the model and initialize the engine
-    model = Patchcore(backbone=backbone, layers=layers, coreset_sampling_ratio=ratio)
+    model = Patchcore(
+        backbone=const.BACKBONE, layers=const.LAYERS, coreset_sampling_ratio=const.RATIO
+    )
     engine = Engine()
 
     # preparation for plot