Opened 18 years ago
Closed 18 years ago
Last modified 18 years ago
#1877 closed enhancement (invalid)
New Commercial Flagger
Reported by: | Robert Tsai <rtsai1111> | Owned by: | cpinkham |
---|---|---|---|
Priority: | minor | Milestone: | unknown |
Component: | mythtv | Version: | head |
Severity: | low | Keywords: | |
Cc: | Ticket locked: | no |
Description
This is a new commercial flagger framework (CommDetector2). It is intended to provide an infrastructure to make adding new analysis strategies easier to add and combine with existing strategies. The initializer registers the various analyzers, and the main loop decodes all the frames and calls each analyzer with that frame.
There is also a brand-new logo detector based on Canny edge detection, with some other building blocks that can be re-used for other strategies (BorderDetector?, PGMConverter). A full description of the logo detector can be seen at http://www.tsaiberspace.net/projects/mythtv/commercials/
Near-term plans:
- Port some of the ClassicCommDetector? strategies over into CommDetector2.
- Add globalsettings for the tunable parameters (currently "const int" in the code, commented with "TUNABLE").
Longer-term plans:
- Provide an interactive way to tune the flagger's parameters. For example, the logo template is currently saved to the filesystem. I'd like to provide a mechanism in the "Edit Recordings" mode to overlay the template with the current frame, with some kind of OSD or something to show why (or why not) a frame was flagged as a commercial (or not).
Attachments (2)
Change History (6)
Changed 18 years ago by
Attachment: | mythtv.commflag.diff added |
---|
comment:1 Changed 18 years ago by
comment:2 Changed 18 years ago by
Resolution: | → invalid |
---|---|
Status: | new → closed |
For now, work will go in to branches/commdetector2
comment:3 Changed 18 years ago by
just wondering if this new infrastructure will more easily allow adding commercial detection via audio fingerprinting?
Should mention a few things: