Mining TV Broadcasts 24/7 for Recurring Video Sequences

Preview:

DESCRIPTION

Mining TV Broadcasts 24/7 for Recurring Video Sequences. Jeong , Dongseok. Before the start – Question. There are two techniques used for Video Fingerprinting : CPF(Color Patches Features) and Gradient Histograms. What is the main idea of these techniques? - PowerPoint PPT Presentation

Citation preview

Mining TV Broadcasts 24/7 for Recurring Video Sequences

Jeong, Dongseok

There are two techniques used for Video Fingerprinting : CPF(Color Patches Features) and Gradient His-tograms. What is the main idea of these techniques?

What methods are used for similar image searching?

Before the start – Question

Detecting recurring video clips in a TV broadcast stream

In this case, to detect recurring commercials in a TV broadcast stream

How can we overcome the storage and computing issue of video data?

The purpose of the paper

IntroductionFingerprinting Video Streams ☜

CPF, Gradient HistogramsExplain own search algorithm ☜Experimental resultsRelated WorkConclusion

The structure of the paper

Color Patches Features(CPF) Each frame is reduced to averages of pixels

Video Fingerprinting

CPF example

Video Fingerprinting

Gradient Histograms Edge-based features

Video Fingerprinting

Gradient Histograms cont.

Video Fingerprinting

Finally, we can measure the distance between two image and

The distance between two sequences , of length L is given by

Video Fingerprinting

Finding similar images to the source video Use inverted index and Locality

Sensitive HashingCompare short clips from each

sourceFinding the start-point and end-point

of repeated sequencesClassifying the repeated video

Searching Algorithm

Clip Length : choose 25 frames

Optimizing some parameters

Minimum Fraction of Matched Frames choose 20%

Optimize some parameters

Maximum Number of Entries in Hash Table choose 100 entries per hash value

Optimize some parameters

Minimum Length of Duplicates Choose 100 frames

Optimize some parameters

Searching for flips with GHs is up to 30% faster than using CPFs

But CPFs are faster to evaluate and need a smaller amount of storage

CPF VS GH

• (a) : Chart TV• (b) : Sky Sports

News

Apply the system to a variety of broadcast stations

Mining Different TV Channels

The remaining false detections are mainly caused by repeated news sto-ries(in ARD : the German public broadcaster)

Gemini is an Indian TV channel – for non-natives what are commercials and what not?

Mining Different TV Channels

There are two techniques used for Video Fingerprinting : CPF(Color Patches Features) and Gradient His-tograms. What is the main idea of these techniques?

What methods are used for similar image searching?

Question Again

Any Question?

The End!

Recommended