Update README.md

README.md

@@ -2,14 +2,13 @@
 View-decoupled Transformer for Person Re-identification under Aerial-ground Camera Network (CVPR'24)
 
 ## [0] Dataset: CARGO
-Dataset Link: [Google Drive](https://drive.google.com/file/d/1yDjyH0VtW7efxP3vgQjIqTx2oafCB67t/view?usp=drive_link)
 
 ### Introduction
 * CARGO is a large-scale aerial-ground person re-identification (AGPReID) dataset, which captured from a synthesized scene in Unity3D.
 * CARGO contains 13 cameras (8 ground and 5 aerial cameras), 5000 person IDs, and 108563 person images.
 * **Camera 1 $\sim$ 5 belong to aerial cameras, and Camera 6 $\sim$ 13 belongs to ground cameras.**
 * In the aerial camera area, two different drone roaming strategies are designed according to the size of the surveillance area. For the small area (left area), we deploy one drone with a $90^\circ$ overhead view, allowing it to move counterclockwise around each street. For a large area (right area), we deploy individual drones on each of the four streets with a $45^\circ\sim60^\circ$ tilt view, allowing them to move back and forth on corresponding streets. 
-
+* Dataset Link: [Google Drive](https://drive.google.com/file/d/1yDjyH0VtW7efxP3vgQjIqTx2oafCB67t/view?usp=drive_link)
 
 ### Setting
 * We split CARGO into the train (51,451 images with 2500 IDs) and test sets (51,024 images with the remaining 2500 IDs) with an almost 1:1 ratio.
@@ -66,4 +65,4 @@ If you find this code useful for your research, please kindly cite the following
     year      = {2024}
 }
 ```
-If you have any question, please feel free to contact me. E-mail: [email protected]
+If you have any question, please feel free to contact me. E-mail: [email protected]