[clug] project, video editing,editing out ads

[steve jenkin]

More search terms:
 “Perceptual Hashing”
 “Video Fingerprinting”
 “Image Deduplication”.

These are active research problems with little Open Source available, at least directly.
There are multiple patents and large, expensive commercial firms doing this work.

I’d argue two things:
 - simple crypto hashes will pick-up Adverts, because they are Bitwise-identical, not perceptual, and
 - there a lot of code out there for image compression. They contain the components to find ‘duplicates’.

I’d use the Brute Force and Ignorance approach first and eliminate bitwise-identical frames/sequences first.
While doing this pass, it’d be wise to do some feature extraction for later use.

HTH
sj


FOSS Project: phash, Perceptual hashing.
<http://www.phash.org>

Big Expensive Product
<http://duplicatevideosearch.com/video-fingerprinting/>

2009, GOOG Open Sources WebP format
<http://www.theregister.co.uk/2011/05/23/webp_format_in_chrome_gmail_picasa/>

2004, BBC Open Sources ‘Dirac’, a wavelet codec.
<http://www.theregister.co.uk/2004/05/10/bbc_develops_codec/>

A 2011-12 paper on the topic from a Stanford course
<http://vision.stanford.edu/teaching/cs231a_autumn1213_internal/project/final/writeup/nondistributable/Wen_Paper.pdf>


> On 15 Apr 2015, at 1:18 pm, steve jenkin <sjenkin at canb.auug.org.au> wrote:
> 
> 
>> On 15 Apr 2015, at 9:57 am, Jason Nielsen <j.lee.nielsen at gmail.com> wrote:
>> 
>> If the station logo / watermark is consistent for the main program and
>> then disappears during the ads you might be able to check for that and
>> drop every frame without it there.
>> 
>> Jason Nielsen
> 
> 
> Jason,
> 
> Good point.
> 
> With digital delivery, we have a very powerful new weapon:
> 

--
Steve Jenkin, IT Systems and Design 
0412 786 915 (+61 412 786 915)
PO Box 48, Kippax ACT 2615, AUSTRALIA

mailto:sjenkin at canb.auug.org.au http://members.tip.net.au/~sjenkin