Update predict_bbox.py

1, added function reverse_resized_rect that resize rect to fit the original image size.
2, switched order of width and height in line 46/53

predict_bbox.py

@@ -23,6 +23,14 @@ def rect_to_bb(rect):
 	# return a tuple of (x, y, w, h)
 	return (x, y, w, h)
 
+def reverse_resized_rect(rect,resize_ratio):
+    l = int(rect.left() / resize_ratio)
+    t = int(rect.top() / resize_ratio)
+    r = int(rect.right() / resize_ratio)
+    b = int(rect.bottom() / resize_ratio)
+    new_rect = dlib.rectangle(l,t,r,b)
+
+    return [l,t,r,b] , new_rect
 
 def detect_face(image_paths,   SAVE_DETECTED_AT, default_max_size=800, size = 300, padding = 0.25):
     cnn_face_detector = dlib.cnn_face_detection_model_v1('dlib_models/mmod_human_face_detector.dat')
@@ -34,12 +42,16 @@ def detect_face(image_paths,   SAVE_DETECTED_AT, default_max_size=800, size = 30
             print('---%d/%d---' %(index, len(image_paths)))
 
         img = dlib.load_rgb_image(image_path)
-        old_width, old_height, _ = img.shape
+
+        old_height, old_width, _ = img.shape
         if old_width > old_height:
-        	new_width, new_height = default_max_size, int(default_max_size * old_height / old_width)
+            resize_ratio = default_max_size / old_width
+            new_width, new_height = default_max_size, int(old_height * resize_ratio)
         else:
-        	new_width, new_height =  int(default_max_size * old_height / old_width), default_max_size
-        img = dlib.resize_image(img, new_width, new_height )
+            resize_ratio = default_max_size / old_height
+            new_width, new_height =  int(old_width * resize_ratio), default_max_size
+        img = dlib.resize_image(img, cols=new_width, rows=new_height)
+
         dets = cnn_face_detector(img, 1)
         num_faces = len(dets)
         if num_faces == 0:
@@ -51,13 +63,15 @@ def detect_face(image_paths,   SAVE_DETECTED_AT, default_max_size=800, size = 30
         for detection in dets:
             rect = detection.rect
             faces.append(sp(img, rect))
-            rects.append(rect)
+            rect_tpl ,rect_in_origin = reverse_resized_rect(rect,resize_ratio)
+            rects.append(rect_in_origin)
         images = dlib.get_face_chips(img, faces, size=size, padding = padding)
         for idx, image in enumerate(images):
             img_name = image_path.split("/")[-1]
             path_sp = img_name.split(".")
             face_name = os.path.join(SAVE_DETECTED_AT,  path_sp[0] + "_" + "face" + str(idx) + "." + path_sp[-1])
-            dlib.save_image(image, face_name)
+            dlib.save_image(image, face_name) 
+
     return rects
 
 def predidct_age_gender_race(save_prediction_at, bboxes, imgs_path = 'cropped_faces/'):
@@ -66,7 +80,8 @@ def predidct_age_gender_race(save_prediction_at, bboxes, imgs_path = 'cropped_fa
 
     model_fair_7 = torchvision.models.resnet34(pretrained=True)
     model_fair_7.fc = nn.Linear(model_fair_7.fc.in_features, 18)
-    model_fair_7.load_state_dict(torch.load('fair_face_models/res34_fair_align_multi_7_20190809.pt'))
+    model_fair_7.load_state_dict(torch.load('fair_face_models/fairface_alldata_20191111.pt'))
+    #model_fair_7.load_state_dict(torch.load('fair_face_models/res34_fair_align_multi_7_20190809.pt'))
     model_fair_7 = model_fair_7.to(device)
     model_fair_7.eval()
 
@@ -220,6 +235,5 @@ def ensure_dir(directory):
     ensure_dir(SAVE_DETECTED_AT)
     imgs = pd.read_csv(args.input_csv)['img_path']
     bboxes = detect_face(imgs, SAVE_DETECTED_AT)
-    print(len(bboxes))
     print("detected faces are saved at ", SAVE_DETECTED_AT)
-    predidct_age_gender_race("test_outputs.csv", bboxes, SAVE_DETECTED_AT)
+    predidct_age_gender_race("test_outputs.csv", bboxes, SAVE_DETECTED_AT)