diff --git a/python/image.py b/python/image.py
index 2346767..7bf12af 100644
--- a/python/image.py
+++ b/python/image.py
@@ -21,7 +21,7 @@
 BAYER_MONO = 'rggb'
 
 
-def load_image(image_path, model=None):
+def load_image(image_path, model=None, debayer=True):
     """Loads and rectifies an image from file.
 
     Args:
@@ -42,7 +42,8 @@ def load_image(image_path, model=None):
         pattern = BAYER_MONO
 
     img = Image.open(image_path)
-    img = demosaic(img, pattern)
+    if debayer:
+        img = demosaic(img, pattern)
     if model:
         img = model.undistort(img)
 
diff --git a/python/play_road_boundary.py b/python/play_road_boundary.py
new file mode 100644
index 0000000..333b538
--- /dev/null
+++ b/python/play_road_boundary.py
@@ -0,0 +1,84 @@
+################################################################################
+#
+# Copyright (c) 2017 University of Oxford
+# Authors:
+#  Daniele De Martini (daniele@robots.ox.ac.uk)
+#
+# This work is licensed under the Creative Commons
+# Attribution-NonCommercial-ShareAlike 4.0 International License.
+# To view a copy of this license, visit
+# http://creativecommons.org/licenses/by-nc-sa/4.0/ or send a letter to
+# Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.
+#
+################################################################################
+
+import argparse
+import cv2
+from pathlib import Path
+from tqdm import tqdm
+import numpy as np
+
+from datetime import datetime as dt
+from road_boundary import load_road_boundary_image, load_road_boundary_mask
+
+
+YELLOW = (255, 255, 0)
+CYAN = (0, 255, 255)
+RED = (255, 0, 0)
+
+
+parser = argparse.ArgumentParser(description='Play back images from a given directory')
+
+parser.add_argument('trial', type=str, help='Directory containing images.')
+parser.add_argument('--camera_id', type=str, default='left', choices=['left', 'right'], help='(optional) Camera ID to display')
+parser.add_argument('--type', type=str, default='uncurated', choices=['uncurated', 'curated'], help='(optional) Curated vs uncurated')
+parser.add_argument('--masks_id', type=str, default='mask', choices=['mask', 'mask_classified'], help='(optional) Masks type to overlay')
+
+parser.add_argument('--save_video',  action='store_true', help='Flag for saving a video')
+parser.add_argument('--save_dir',  type=str, help='Where to save the images')
+
+args = parser.parse_args()
+
+image_path = Path(args.trial) / args.type / args.camera_id / 'rgb'
+mask_path = Path(args.trial) / args.type / args.camera_id / args.masks_id
+
+assert image_path.exists(), f'Image path {image_path} does not exist'
+assert mask_path.exists(), f'Mask path {mask_path} does not exist'
+
+images = sorted(image_path.glob('*.png'))
+masks = sorted(mask_path.glob('*.png'))
+
+fname = f"{args.trial}_{args.camera_id}_{args.type}_{args.masks_id}"
+
+initialised = False
+for image, mask in tqdm(zip(images, masks)):
+    image = load_road_boundary_image(str(image))
+    mask = load_road_boundary_image(str(mask))
+
+    kernel = np.ones((5, 5), 'uint8')
+    mask = cv2.dilate(mask, kernel, iterations=1)
+
+    if args.masks_id == 'mask':
+        image[mask > 0] = YELLOW
+    else:
+        image[mask == 1] = CYAN
+        image[mask == 2] = RED
+
+    image_ = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
+    cv2.imshow('Video', image_)
+    if cv2.waitKey(1) & 0xFF == ord('q'):
+        break
+
+    if args.save_video:
+        if not initialised:
+            framesize = (image.shape[1], image.shape[0])
+            out = cv2.VideoWriter(str(Path(args.save_dir) / f'{fname}.avi'), \
+                                  cv2.VideoWriter_fourcc(*'MPEG'),
+                                  20, framesize, True)
+            initialised = True
+
+            cv2.imwrite(str(Path(args.save_dir) / f'{fname}.jpg'), image_)
+
+        out.write(image_)
+
+out.release()
diff --git a/python/road_boundary.py b/python/road_boundary.py
new file mode 100644
index 0000000..453649d
--- /dev/null
+++ b/python/road_boundary.py
@@ -0,0 +1,25 @@
+################################################################################
+#
+# Copyright (c) 2017 University of Oxford
+# Authors:
+#  Daniele De Martini (daniele@robots.ox.ac.uk)
+#
+# This work is licensed under the Creative Commons
+# Attribution-NonCommercial-ShareAlike 4.0 International License.
+# To view a copy of this license, visit
+# http://creativecommons.org/licenses/by-nc-sa/4.0/ or send a letter to
+# Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.
+#
+################################################################################
+
+import numpy as np
+from PIL import Image
+
+
+def load_road_boundary_image(image_path):
+    img = Image.open(image_path)
+    return np.array(img).astype(np.uint8)
+
+
+def load_road_boundary_mask(mask_path, model):
+    return load_road_boundary_image(mask_path)