VCSL

VCSL copied to clipboard

The paper provides benchmark tests under different data distributions. I would like to ask how the "Benchmark F-score results at different segment durations" are obtained, because each pair of videos...

loujinming147

Video Quality(360p/720p/1080p) To Choose

1

Hi, I was wondering do you control the video quality(e.g. 360p/720p/1080p) when downloading the videos from youtube/bilibili, since tools like yt-dlp will automatically choose the best quality if using default...

foollinkie

Did not reproduce the results on git

1

Thanks for the excellent work. Following the scripts on the git (ISC feature + DTW), I got the following results Recall: 52.35% Precision: 42.44% F1 score: 50.32% The reproduced results...

KT27-A

Some of bilibili videos are not available

1

When I use youtube-dl to download videos, I found some of bilibili videos are not available(such as https://www.bilibili.com/video/BV1TB4y1P7z8 ).Do you have any backups?

buzepeng

Some labels are duplicate

1

Thanks for your great work on VCSL dataset! When checking test pair in `split_meta_pairs.json`, we found about 1% of the labels are self-duplicate (e.g. `e45b9eec5ea54e00a2d6e6689cd5fe92-e45b9eec5ea54e00a2d6e6689cd5fe92`, `dd44c4be9fdc4c95bdf075e7e756294e-dd44c4be9fdc4c95bdf075e7e756294e`) or query-reference reverse duplicate...

xuan97916

Could you please provide the training code?

1

Nice work! I am interested in this work. If availabel, could you please provide the training code?

chenqz1998

More Feature

1

Could you integrate our [3st Place Solution of the Facebook AI Image Similarity Challenge](https://github.com/WangWenhao0716/ISC-Track2-Submission) into your repo? This is a very easy and well-generalized version of our [1st Place Solution...

WangWenhao0716

missing test set for VCSL_Dataset

1

Thanks for your nice work! But I can't download all the test set, there are only 37087 of the 55530 pairs were downloaded. In the case, how can I use...

lst17640097780