Page 1 of 1

MPEG7 Descriptors

PostPosted: Wed Sep 21, 2011 7:44 am
by ahyeekvision
Any plan in future for adding in MPEG7 descriptors in AForge.NET?

Re: MPEG7 Descriptors

PostPosted: Wed Sep 21, 2011 8:31 am
by andrew.kirillov
No, no plans for this in the coming future. Plus there is not much sense of adding this to the framework, which is mostly targeted to computer vision and surrounding areas.
"MPEG-7 is intended to provide complementary functionality to the previous MPEG standards, representing information about the content, not the content itself ..."

P.S. By the way ... Why do you ask about AForge.NET framework in IPLab forum? Is it so hard to select proper board?

Re: MPEG7 Descriptors

PostPosted: Wed Sep 21, 2011 9:41 am
by ahyeekvision
Aha. Thought the extraction of feature descriptors from an image can be considered one of the image processing methods to be discussed in IPLab. Hence, such image feature extraction processing method can be considered to be added in AForge.NET framework.

I am not sure. If you consider this topic is more suitable to park under AForge.NET board, can you please kindly move it to there :).
Thanks.

By the way, don't you think feature extraction algorithms for extracting descriptors such as SIFT, SURF (http://en.wikipedia.org/wiki/SURF) and etc are important and useful to be considered in the framework?

In fact, MPEG-7 visual descriptors are also being widely researched and published since year 2010.
Anyway, this is just some suggestions for discussion :)

Re: MPEG7 Descriptors

PostPosted: Wed Sep 21, 2011 10:02 am
by andrew.kirillov
ahyeekvision wrote:Thought the extraction of feature descriptors from an image can be considered one of the image processing methods to be discussed in IPLab. Hence, such image feature extraction processing method can be considered to be added in AForge.NET framework.

A bit confused here. IPLab is just an application for applying different image processing routines to still images. It has nothing to do with video. Plus you said in the very beginning: " ... adding in MPEG7 descriptors in AForge.NET?"

ahyeekvision wrote:By the way, don't you think feature extraction algorithms for extracting descriptors such as SIFT, SURF (http://en.wikipedia.org/wiki/SURF) and etc are important and useful to be considered in the framework?

SURF/SIFT features seems to be a bit different from MPEG7 Descriptorsm ... If SURF is targeted to features like edges, corners, etc., then MPEG7 descriptors provide information about audio/video content, which could be song's lyrics, description of currently shown scene in video, etc., etc. So if we speak about content description, then this is not a priority for the framework for now. But if we speak about SURF like things, then yes - there is interest to it.

Re: MPEG7 Descriptors

PostPosted: Wed Sep 21, 2011 10:23 am
by ahyeekvision
Ops.
I think you have understand the MPEG-7 Descriptors as some kind of "metadata" to index a video / audio, but it is not.
It is something like visual descriptors. It's not only able to applied to video, but also still image for CBIR purpose.
The descriptors talking about things like: Scalable Color, Color Layout , Dominant Color(s) and etc.

More info: http://savvash.blogspot.com/p/mpeg-7-descriptors.html
This web I referred write quite a number of details about the descriptor functions.

Hope clarify.

Re: MPEG7 Descriptors

PostPosted: Wed Sep 21, 2011 10:38 am
by andrew.kirillov
Yes, this one looks more interesting and something to think about ;) The only issues is lack of time to fit everything interesting into the framework :( So for now MPEG7 will stay out of scope till better time comes. And there is C# code already exists for it anyway ...