Add haarscade support for mp4

const.py

@@ -6,8 +6,10 @@
 TRAIN_TF_RECORD_FILE = os.path.join(INPUT_DIR, "train.tfrecord")
 VAL_TF_RECORD_FILE = os.path.join(INPUT_DIR, "val.tfrecord")
 
-YOLO_SIZE = 352
-YOLO_TARGET = 11
+# YOLO_SIZE = 352
+# YOLO_TARGET = 11
+YOLO_SIZE = 288
+YOLO_TARGET = 9
 
 CORRECTOR_SIZE = 50
 

detector.py

@@ -61,7 +61,7 @@ def load_model_pb(self, model_dir):
             self.sess = tf.Session()
 
 
-    def predict(self, frame, thresh=0.8):
+    def predict(self, frame, thresh=0.9):
         input_img = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
         input_img = cv2.resize(input_img, (YOLO_SIZE, YOLO_SIZE)) / 255.
         input_img = np.reshape(input_img, [1, YOLO_SIZE, YOLO_SIZE, 3])

haarcascade_video_client.py

@@ -0,0 +1,68 @@
+import cv2
+import sys
+import time
+
+from detector import FaceDetector
+from utils import annotate_image
+
+
+def run(feed):
+    cascPath = "./env/lib/python3.6/site-packages/cv2/data/haarcascade_frontalface_default.xml"
+    faceCascade = cv2.CascadeClassifier(cascPath)
+
+    if feed is None:
+        # From webcam
+        cap = cv2.VideoCapture(0)
+    else:
+        cap = cv2.VideoCapture(feed)
+
+    # Get current width of frame
+    width = cap.get(cv2.CAP_PROP_FRAME_WIDTH)   # float
+    # Get current height of frame
+    height = cap.get(cv2.CAP_PROP_FRAME_HEIGHT) # float
+    fps = cap.get(cv2.CAP_PROP_FPS) # float
+
+
+    # Define the codec and create VideoWriter object
+    # fourcc = cv2.Video.CV_FOURCC(*'X264')
+    fourcc = cv2.VideoWriter_fourcc(*'MJPG')
+    out = cv2.VideoWriter("output.avi",fourcc, fps, (int(width),int(height)))
+
+    now = time.time()
+    while(cap.isOpened()):
+        now = time.time()
+
+        ret, frame = cap.read()
+
+        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
+
+        now = time.time()
+
+        faces = faceCascade.detectMultiScale(
+            gray,
+            scaleFactor=1.1,
+            minNeighbors=5,
+            minSize=(30, 30),
+            flags=cv2.CASCADE_SCALE_IMAGE
+        )
+
+        # Draw a rectangle around the faces
+        for (x, y, w, h) in faces:
+            cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 0), 2)
+
+        out.write(frame)
+
+        # Display the resulting frame
+        # cv2.imshow('frame', ann_frame)
+        print("FPS: {:0.2f}".format(1 / (time.time() - now)))
+        if cv2.waitKey(1) & 0xFF == ord('q'):
+            break
+
+    # When everything done, release the capture
+    cap.release()
+    cv2.destroyAllWindows()
+
+
+if __name__ == "__main__":
+    feed = None if len(sys.argv) == 1 else sys.argv[1]
+    run(feed)

video_client.py

@@ -5,8 +5,8 @@
 from detector import FaceDetector
 from utils import annotate_image
 
-YOLO_MODELS_DIR = "/home/ivanitz/yolo-face/models/"
-CORRECTOR_MODELS_DIR = "/home/ivanitz/face-correction/models/"
+YOLO_MODELS_DIR = "../yolo-face-artifacts/run8/models/"
+CORRECTOR_MODELS_DIR = "../fine-tuned-face/models/"
 
 
 def run(feed):
@@ -33,20 +33,22 @@ def run(feed):
 
     now = time.time()
     while(cap.isOpened()):
+        now = time.time()
         # Capture frame-by-frame
         ret, frame = cap.read()
 
-        now = time.time()
+        # now = time.time()
         bboxes = face_detector.predict(frame)
-        print("FPS: {:0.2f}".format(1 / (time.time() - now)))
+        # print("FPS: {:0.2f}".format(1 / (time.time() - now)))
         ann_frame = annotate_image(frame, bboxes)
 
         out.write(ann_frame)
 
         # Display the resulting frame
         # cv2.imshow('frame', ann_frame)
-        # if cv2.waitKey(1) & 0xFF == ord('q'):
-        #     break
+        print("FPS: {:0.2f}".format(1 / (time.time() - now)))
+        if cv2.waitKey(1) & 0xFF == ord('q'):
+            break
 
     # When everything done, release the capture
     cap.release()