Update README.md
Browse files
README.md
CHANGED
|
@@ -1,18 +1,30 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
size_categories:
|
| 4 |
+
- 1K<n<10K
|
| 5 |
+
---
|
| 6 |
+
|
| 7 |
+
## AVE Dataset
|
| 8 |
+
|
| 9 |
+
The original AVE dataset ported from the (GitHub)[https://github.com/YapengTian/AVE-ECCV18/blob/master/supervised_main.py]
|
| 10 |
+
|
| 11 |
+
## Notes
|
| 12 |
+
|
| 13 |
+
- One video may contain different audio-visual events, so the total number of videos is not 4143.
|
| 14 |
+
- annotations.txt: annotations of AVE dataset. For each sample, you can find its event catergory, YouTube ID, Quality (all good, means that it contains an AVE), start time of an audio-visual event, the end time of an audio-visual event.
|
| 15 |
+
- train/val/test-Set.txt: training/validation/testing set used in our ECCV paper.
|
| 16 |
+
|
| 17 |
+
## Citation
|
| 18 |
+
|
| 19 |
+
If you used the AVE dataset, please consider citing the original paper:
|
| 20 |
+
|
| 21 |
+
```
|
| 22 |
+
@inproceedings{TianECCV2018,
|
| 23 |
+
title={Audio-Visual Event Localization in Unconstrained Videos},
|
| 24 |
+
author={Tian, Yapeng and Shi, Jing and Li, Bochen and Duan, Zhiyao and Xu, Chenliang},
|
| 25 |
+
editor="Ferrari, Vittorio and Hebert, Martial and Sminchisescu, Cristian and Weiss, Yair",
|
| 26 |
+
booktitle="Computer Vision -- ECCV 2018",
|
| 27 |
+
year="2018",
|
| 28 |
+
publisher="Springer",
|
| 29 |
+
}
|
| 30 |
+
```
|