[mythtv] A thought on Commercial Flagging -- Long (And rambling)

[Lucas Meijer]

Robert Johnston wrote:

> What I'm suggesting is that rather than having to keep adding bits to
> the commercial flagging routines every time the broadcasters change
> their tack, which ties up development time that could be used on other
> things, we have a "Learning" system that will adapt "on the box" to
> improve commercial detection in a continuing and ongoing way, without
> having to rewrite, rebuild, or recompile.

It looks like you have a very skewed idea of the development and testing 
time required for such a system.

> This is what I'm talking about. I'm suggesting using a "Learning"
> algorithm for commercial flagging to work from, which will learn and
> constantly improve itself based on user's input, tailoring itself to
> their viewing preferences and wants.

algorithms don't "just learn". You need to provide it with sensible 
input. Like the amount of camera switches, wether or not there is a 
logo, sound levels etc... Just like the current system needs those 
input. Neural nets or baysian filters aren't magical. They need the same 
inputs as the current system, and I'd bet both nets and baysian filters 
will be greatly outperformed by a handmade ruleset.

the problem is getting the the input parameters. once the detection of 
those works better, commercial flagging gets much better. while it 
doesn't work, there ain't no funky learningalgorithm thats suddenly 
going to work well.

> I agree that we need more eyes on this area, more people helping out.
> I'm just offering a suggestion as to how we can "Solve" the commercial
> flagging problem once and for all.

Sure thing. I'm just trying to make sure the people who might be 
listening and thinking "hey, I could help with that" realize that there 
are other areas that need help that have a much better value/timespent 
ratio.

Bye, Lucas