From: Leandro Santiago <leandrosansilva@gmail.com>
To: FFmpeg development discussions and patches
<ffmpeg-devel@ffmpeg.org>,
Michael Niedermayer <michael@niedermayer.cc>
Subject: Re: [FFmpeg-devel] I've written a filter in Rust
Date: Fri, 21 Feb 2025 08:56:04 +0100
Message-ID: <c99ee333-4a14-4599-a804-9ee82feed537@gmail.com> (raw)
In-Reply-To: <20250220224924.GL4991@pb2>
On 2/20/25 23:49, Michael Niedermayer wrote:
> Hi
>
> On Thu, Feb 20, 2025 at 02:06:47PM +0100, Leandro Santiago wrote:
>> [insert meme here]
> [...]
>> I also recorded a video showing the filter in action [7].
> [...
>> [7] https://youtu.be/U_y4-NnaINg
> cool, it doesnt detect everyone though
Yes, this has two reasons: one is the quality of the detection depends on the model used. I use in the demo a yolov4-tiny that is very fast but not really accurate, leading to many people not being tracked. There can be bugs in the track_sort code as well.
The other reason is that the `track_sort` at the moment ignores all the detected boxes which have a position, but no area (they seem to be detected as simple points by `dnn_detect`), which are useless for now, as the SORT implementation at the moment depends only the boxes dimensions for computing IOU (intersection over union).
I could think though on extending the code to "fake an area" of a box when it's unknown, which might help to improve the tracking.
>
> also i think this shows how useful a plugin framework would be for ffmpeg
>
> with plugins everyone could use,test and contribute to this today.
> without plugins, this needs to be merged in ffmpeg git master. (which
> will take some time i suspect)
Yes, it could even open space for webassembly based plugins, which would be less of a hassle regarding ABI and potentially be cross platform, although I foresee some challenges on getting wasm runtime to share data with the host without copying, especially if the data in the GPU.
>
> Rust support in FFmpeg is an interresting idea though, iam in favor of
> supporting more "safe" languages
Great to hear that!
>
> thx
>
> [...]
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
next prev parent reply other threads:[~2025-02-21 7:56 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-02-20 13:06 Leandro Santiago
2025-02-20 16:20 ` Leandro Santiago
2025-02-20 22:49 ` Michael Niedermayer
2025-02-21 7:56 ` Leandro Santiago [this message]
2025-02-21 9:01 ` Tomas Härdin
2025-02-21 9:21 ` Soft Works
2025-02-21 13:21 ` Michael Niedermayer
2025-02-21 14:30 ` Soft Works
2025-02-21 14:53 ` Kieran Kunhya via ffmpeg-devel
2025-02-21 15:02 ` Soft Works
2025-02-21 19:27 ` Kieran Kunhya via ffmpeg-devel
2025-02-21 20:10 ` Soft Works
2025-02-26 13:50 ` Tomas Härdin
2025-02-26 14:18 ` Zhao Zhili
2025-02-26 15:32 ` Rémi Denis-Courmont
2025-02-26 16:03 ` Zhao Zhili
2025-02-26 16:25 ` martin schitter
2025-02-26 14:07 ` Nicolas George
2025-02-26 16:35 ` Soft Works
2025-02-27 22:40 ` Michael Niedermayer
2025-02-21 16:39 ` Stephen Hutchinson
2025-02-26 14:25 ` Vittorio Giovara
2025-02-21 13:18 ` Lynne
2025-02-21 13:44 ` Kieran Kunhya via ffmpeg-devel
2025-02-21 18:02 ` Tomas Härdin
2025-02-22 12:57 ` Rémi Denis-Courmont
2025-02-23 21:30 ` Tomas Härdin
2025-02-23 21:51 ` Michael Niedermayer
2025-02-26 14:11 ` Tomas Härdin
2025-02-27 21:01 ` Michael Niedermayer
2025-02-28 1:57 ` Pavel Koshevoy
2025-02-28 15:35 ` Rémi Denis-Courmont
2025-02-24 14:51 ` Rémi Denis-Courmont
2025-02-26 14:34 ` Tomas Härdin
2025-02-26 15:13 ` Leandro Santiago
2025-02-22 12:49 ` Rémi Denis-Courmont
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=c99ee333-4a14-4599-a804-9ee82feed537@gmail.com \
--to=leandrosansilva@gmail.com \
--cc=ffmpeg-devel@ffmpeg.org \
--cc=michael@niedermayer.cc \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
This inbox may be cloned and mirrored by anyone:
git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git
# If you have public-inbox 1.1+ installed, you may
# initialize and index your mirror using the following commands:
public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
ffmpegdev@gitmailbox.com
public-inbox-index ffmpegdev
Example config snippet for mirrors.
AGPL code for this site: git clone https://public-inbox.org/public-inbox.git