set left and right eye locations with respect to the person instead of observer

deepface/detectors/Dlib.py

@@ -88,11 +88,11 @@ class DlibClient(Detector):
 
                 shape = self.model["sp"](img, detection)
 
-                left_eye = (
+                right_eye = (
                     int((shape.part(2).x + shape.part(3).x) // 2),
                     int((shape.part(2).y + shape.part(3).y) // 2),
                 )
-                right_eye = (
+                left_eye = (
                     int((shape.part(0).x + shape.part(1).x) // 2),
                     int((shape.part(0).y + shape.part(1).y) // 2),
                 )

deepface/detectors/FastMtCnn.py

@@ -34,8 +34,8 @@ class FastMtCnnClient(Detector):
         ):
             for regions, confidence, eyes in zip(*detections):
                 x, y, w, h = xyxy_to_xywh(regions)
-                left_eye = eyes[0]
+                right_eye = eyes[0]
-                right_eye = eyes[1]
+                left_eye = eyes[1]
 
                 left_eye = tuple(int(i) for i in left_eye)
                 right_eye = tuple(int(i) for i in right_eye)

deepface/detectors/MediaPipe.py

@@ -61,8 +61,8 @@ class MediaPipeClient(Detector):
             y = int(bounding_box.ymin * img_height)
             h = int(bounding_box.height * img_height)
 
-            left_eye = (int(landmarks[0].x * img_width), int(landmarks[0].y * img_height))
+            right_eye = (int(landmarks[0].x * img_width), int(landmarks[0].y * img_height))
-            right_eye = (int(landmarks[1].x * img_width), int(landmarks[1].y * img_height))
+            left_eye = (int(landmarks[1].x * img_width), int(landmarks[1].y * img_height))
             # nose = (int(landmarks[2].x * img_width), int(landmarks[2].y * img_height))
             # mouth = (int(landmarks[3].x * img_width), int(landmarks[3].y * img_height))
             # right_ear = (int(landmarks[4].x * img_width), int(landmarks[4].y * img_height))

deepface/detectors/MtCnn.py

@@ -35,8 +35,10 @@ class MtCnnClient(Detector):
             for current_detection in detections:
                 x, y, w, h = current_detection["box"]
                 confidence = current_detection["confidence"]
-                left_eye = current_detection["keypoints"]["left_eye"]
+                # mtcnn detector assigns left eye with respect to the observer
-                right_eye = current_detection["keypoints"]["right_eye"]
+                # but we are setting it with respect to the person itself
+                left_eye = current_detection["keypoints"]["right_eye"]
+                right_eye = current_detection["keypoints"]["left_eye"]
 
                 facial_area = FacialAreaRegion(
                     x=x,

deepface/detectors/OpenCv.py

@@ -112,15 +112,18 @@ class OpenCvClient(Detector):
             eye_2 = eyes[1]
 
             if eye_1[0] < eye_2[0]:
-                left_eye = eye_1
-                right_eye = eye_2
-            else:
-                left_eye = eye_2
                 right_eye = eye_1
+                left_eye = eye_2
+            else:
+                right_eye = eye_2
+                left_eye = eye_1
 
             # -----------------------
             # find center of eyes
-            left_eye = (int(left_eye[0] + (left_eye[2] / 2)), int(left_eye[1] + (left_eye[3] / 2)))
+            left_eye = (
+                int(left_eye[0] + (left_eye[2] / 2)),
+                int(left_eye[1] + (left_eye[3] / 2)),
+            )
             right_eye = (
                 int(right_eye[0] + (right_eye[2] / 2)),
                 int(right_eye[1] + (right_eye[3] / 2)),

deepface/detectors/RetinaFace.py

@@ -34,9 +34,9 @@ class RetinaFaceClient(Detector):
             x = detection[0]
             w = detection[2] - x
 
-            # notice that these must be inverse for retinaface
+            # retinaface sets left and right eyes with respect to the person
-            left_eye = identity["landmarks"]["right_eye"]
+            left_eye = identity["landmarks"]["left_eye"]
-            right_eye = identity["landmarks"]["left_eye"]
+            right_eye = identity["landmarks"]["right_eye"]
 
             # eyes are list of float, need to cast them tuple of int
             left_eye = tuple(int(i) for i in left_eye)

deepface/detectors/Yolo.py

@@ -81,10 +81,10 @@ class YoloClient(Detector):
             x, y, w, h = result.boxes.xywh.tolist()[0]
             confidence = result.boxes.conf.tolist()[0]
 
-            # left_eye_conf = result.keypoints.conf[0][0]
+            # right_eye_conf = result.keypoints.conf[0][0]
-            # right_eye_conf = result.keypoints.conf[0][1]
+            # left_eye_conf = result.keypoints.conf[0][1]
-            left_eye = result.keypoints.xy[0][0].tolist()
+            right_eye = result.keypoints.xy[0][0].tolist()
-            right_eye = result.keypoints.xy[0][1].tolist()
+            left_eye = result.keypoints.xy[0][1].tolist()
 
             # eyes are list of float, need to cast them tuple of int
             left_eye = tuple(int(i) for i in left_eye)

deepface/detectors/YuNet.py

@@ -99,7 +99,7 @@ class YuNetClient(Detector):
             {x, y}_{re, le, nt, rcm, lcm} stands for the coordinates of right eye,
             left eye, nose tip, the right corner and left corner of the mouth respectively.
             """
-            (x, y, w, h, x_re, y_re, x_le, y_le) = list(map(int, face[:8]))
+            (x, y, w, h, x_le, y_le, x_re, y_re) = list(map(int, face[:8]))
 
             # YuNet returns negative coordinates if it thinks part of the detected face
             # is outside the frame.