Performance Tweaks vorgenommen

projectmood/face_detect.py

@@ -3,10 +3,10 @@ This module contains face detections functions.
 """
 import cv2
 
-faceDet = cv2.CascadeClassifier("Haarcascade\\haarcascade_frontalface_default.xml")
-faceDet_two = cv2.CascadeClassifier("Haarcascade\\haarcascade_frontalface_alt2.xml")
-faceDet_three = cv2.CascadeClassifier("Haarcascade\\haarcascade_frontalface_alt.xml")
-faceDet_four = cv2.CascadeClassifier("Haarcascade\\haarcascade_frontalface_alt_tree.xml")
+faceDet = cv2.CascadeClassifier('Haarcascade/haarcascade_frontalface_default.xml')
+faceDet_two = cv2.CascadeClassifier('Haarcascade/haarcascade_frontalface_alt2.xml')
+faceDet_three = cv2.CascadeClassifier('Haarcascade/haarcascade_frontalface_alt.xml')
+faceDet_four = cv2.CascadeClassifier('Haarcascade/haarcascade_frontalface_alt_tree.xml')
 
 
 def find_faces(image):
@@ -24,23 +24,28 @@ def _normalize_face(face):
 
 
 def _locate_faces(image):
-    face_one = faceDet.detectMultiScale(image, scaleFactor=1.1, minNeighbors=15, minSize=(70, 70),
+    face_one = faceDet.detectMultiScale(image, scaleFactor=2, minNeighbors=4, minSize=(50, 50),
                                         flags=cv2.CASCADE_SCALE_IMAGE)
-    face_two = faceDet_two.detectMultiScale(image, scaleFactor=1.1, minNeighbors=15, minSize=(70, 70),
+
+    if len(face_one) == 1:
+        return face_one
+
+    face_two = faceDet_two.detectMultiScale(image, scaleFactor=2, minNeighbors=4, minSize=(50, 50),
                                             flags=cv2.CASCADE_SCALE_IMAGE)
-    face_three = faceDet_three.detectMultiScale(image, scaleFactor=1.1, minNeighbors=15, minSize=(70, 70),
+
+    if len(face_two) == 1:
+        return face_two
+
+    face_three = faceDet_three.detectMultiScale(image, scaleFactor=2, minNeighbors=4, minSize=(50, 50),
                                                 flags=cv2.CASCADE_SCALE_IMAGE)
-    face_four = faceDet_four.detectMultiScale(image, scaleFactor=1.1, minNeighbors=15, minSize=(70, 70),
+
+    if len(face_three) == 1:
+        return face_three
+
+    face_four = faceDet_four.detectMultiScale(image, scaleFactor=2, minNeighbors=4, minSize=(50, 50),
                                               flags=cv2.CASCADE_SCALE_IMAGE)
-    if len(face_one) == 1:
-        facefeatures = face_one
-    elif len(face_two) == 1:
-        facefeatures = face_two
-    elif len(face_three) == 1:
-        facefeatures = face_three
-    elif len(face_four) == 1:
-        facefeatures = face_four
-    else:
-        facefeatures = ""
-
-    return facefeatures
+
+    if len(face_four) == 1:
+        return face_four
+
+    return ''

projectmood/webcam.py

@@ -40,15 +40,17 @@ def show_webcam_and_run(model, emoticons, window_size=None, window_name='webcam'
     if vc.isOpened():
         read_value, webcam_image = vc.read()
     else:
-        print("webcam not found")
+        print("Webcam nicht gefunden")
         return
 
     while read_value:
         for normalized_face, (x, y, w, h) in find_faces(webcam_image):
             prediction = model.predict(normalized_face)  # do prediction
+            cv2.imshow("Face", normalized_face)
 
             image_to_draw = emoticons[(prediction[0])]
-            draw_with_alpha(webcam_image, image_to_draw, (x-150, y-50, w-150, h-150))
+            if x - 150 > 0 and y - 50 > 0 and w - 150 > 0 and h - 150 > 0:
+                draw_with_alpha(webcam_image, image_to_draw, (x-150, y-50, w-150, h-150))
 
         cv2.imshow(window_name, webcam_image)
         read_value, webcam_image = vc.read()
@@ -67,7 +69,7 @@ if __name__ == '__main__':
 
     # load mode
     fisher_face = cv2.face.FisherFaceRecognizer_create()
-    fisher_face.read("Basis_data\\models\\detection_model.xml")
+    fisher_face.read('Basis_data/models/detection_model.xml')
     # use learnt model
     window_name = 'WEBCAM (press ESC to exit)'
-show_webcam_and_run(fisher_face, emoticons, window_size=(800, 800), window_name=window_name, update_time=10)
+show_webcam_and_run(fisher_face, emoticons, window_size=(800, 800), window_name=window_name, update_time=1)