remove watermark removal demos.

README.md

@@ -30,7 +30,7 @@
         <a href="https://arxiv.org/abs/2309.03897" target='_blank'>
         <img src="https://img.shields.io/badge/arXiv-2309.03897-b31b1b.svg">
         </a>
-        <a href="https://youtu.be/Cc89WF-2zz0" target='_blank'>
+        <a href="https://youtu.be/92EHfgCO5-Q" target='_blank'>
         <img src="https://img.shields.io/badge/Demo%20Video-%23FF0000.svg?logo=YouTube&logoColor=white">
         </a>
         <img src="https://api.infinitescript.com/badgen/count?name=sczhou/ProPainter">
@@ -48,7 +48,7 @@
 
 
 ## Update
-- **2023.09.24**: We remove the watermark removal demo officially to prevent the misuse of our work for unethical purposes.
+- **2023.09.24**: We remove the watermark removal demos officially to prevent the misuse of our work for unethical purposes.
 - **2023.09.21**: Add features for memory-efficient inference. Check our [GPU memory](https://github.com/sczhou/ProPainter#-memory-efficient-inference) requirements. 🚀
 - **2023.09.07**: Our code and model are publicly available. 🐳
 - **2023.09.01**: This repo is created.
@@ -83,6 +83,14 @@
       <img src="assets/video_completion2.gif">
    </td>
 </tr>
+<tr>
+   <td> 
+      <img src="assets/video_completion3.gif">
+   </td>
+   <td> 
+      <img src="assets/video_completion4.gif">
+   </td>
+</tr>
 </table>
 
 
@@ -136,7 +144,7 @@ Run the following commands to try it out:
 # The first example (object removal)
 python inference_propainter.py --video inputs/object_removal/bmx-trees --mask inputs/object_removal/bmx-trees_mask 
 # The second example (video completion)
-python inference_propainter.py --video inputs/video_completion/running_car.mp4 --mask inputs/video_completion/mask.png
+python inference_propainter.py --video inputs/video_completion/running_car.mp4 --mask inputs/video_completion/mask_square.png --height 240 --width 432
 ```
 
 The results will be saved in the `results` folder.
@@ -145,7 +153,7 @@ To test your own videos, please prepare the input `mp4 video` (or `split frames`
 If you want to specify the video resolution for processing or avoid running out of memory, you can set the video size of `--width` and `--height`:
 ```shell
 # process a 576x320 video; set --fp16 to use fp16 (half precision) during inference.
-python inference_propainter.py --video inputs/video_completion/running_car.mp4 --mask inputs/video_completion/mask.png --height 320 --width 576 --fp16
+python inference_propainter.py --video inputs/video_completion/running_car.mp4 --mask inputs/video_completion/mask_square.png --height 320 --width 576 --fp16
 ```
 
 ### 🚀 Memory-efficient inference

inference_propainter.py

@@ -165,7 +165,7 @@ def get_ref_index(mid_neighbor_id, neighbor_ids, length, ref_stride=10, ref_num=
     else:
         start_idx = max(0, mid_neighbor_id - ref_stride * (ref_num // 2))
         end_idx = min(length, mid_neighbor_id + ref_stride * (ref_num // 2))
-        for i in range(start_idx, end_idx + 1, ref_stride):
+        for i in range(start_idx, end_idx, ref_stride):
             if i not in neighbor_ids:
                 if len(ref_index) > ref_num:
                     break
@@ -254,6 +254,7 @@ if __name__ == '__main__':
         green = np.zeros([h, w, 3]) 
         green[:,:,1] = 255
         alpha = 0.6
+        # alpha = 1.0
         fuse_img = (1-alpha)*img + alpha*green
         fuse_img = mask_ * fuse_img + (1-mask_)*img
         masked_frame_for_save.append(fuse_img.astype(np.uint8))