How to detect the logo from the image?

I want to detect the logo from the image but the image clarity is not good.Here is the image.
Please help me to extract the logo.

Risposte (2)

Image Analyst
Image Analyst il 26 Dic 2013
It looks like it says Animal Planet. Use imrect() or rbbox() to extract it manually, or use regular indexing if you know the rows and columns it occupies.

10 Commenti

Yes i know the coordinates but can you please tell me how to match that logo with original Animal Planet logo.
What does match mean to you?
Naishil shah
Naishil shah il 26 Dic 2013
Modificato: Naishil shah il 26 Dic 2013
here i want to match this logo
with the original logo
I assume what you intend to say (but didn't) is that you want to find that logo automatically in the image no matter where it appears, what color it is, or how it is blended in with clutter in the same place. That's pretty difficult, like trying to decode a captcha.
What's the use case for this? It seems like it would be a copyright violation. I don't want to give advice to you if it is just for you to try to break the law and alter the copyrighted video. Do you have legal permission to do this? What will you do with the videos if you were able to remove the logo?
Naishil shah
Naishil shah il 26 Dic 2013
Modificato: Naishil shah il 26 Dic 2013
See its my project.I just want to match the logo and it will automatically give the answer.Do you know about the TV rating system?This logo matching is a part of that system.So is it a copyright violation???If yes then its OK. And i am doing this operations in images only.This is my dissertation work.I just want to know that which operation should i do to enhance that image. Thank you.
You haven't given an answer as to why you want to do it. I know it's your project but who is sponsoring it and why? Is it just like homework? Maybe you'd get an exemption because it's an educational research thing - I don't know. I mean if the purpose is to just get rid of the annoying video bugs (yes, that's what they're called) why not just crop the frames? I think any other method will be too complicated for you, and maybe for me as well. I can think of one method that might work using edge detection and Hausdorf distances but even that won't get it for every frame. Your best bet is to assume it's going to be in one of the corners and look for bimodal histograms which could indicate a bug over a uniform background. Then do a normalized cross correlation or the edge/Hausdorf method to locate it.
See here for more on video bugs:
Naishil shah
Naishil shah il 26 Dic 2013
Modificato: Naishil shah il 26 Dic 2013
Yes its just like homework. Nobody is sponsoring me for this.I want to select my thesis topic and i have found that for TV rating system they are doing this operation.In this they just match the logo and find out the name of the logo.After finding the name they can rate that Channel easily.I think for this first i have to increase color saturation.
Good luck trying some of the things I suggested, and on developing your own algorithms. For your thesis you'll most likely have to use non-copyrighted videos, but you can create videos with bugs easily yourself with DVD authoring programs like Pinnacle Studio and others. I would first start with non-transparent bugs over uniform backgrounds, then graduate to cluttered backgrounds and transparent (edge only) bugs as you become more experienced.
OK,i will make my own video and thank you for your suggestion but if i am using the bugs which is transparent then again it will become tough to recognize.Can i use thresholding technique to detect the transparent logo???
For transparent ones, where you see just the edges of the bug, then you'll probably have to do an edge detection and see if the shape of the edges matches the shape of the edges of a solid logo.

Accedi per commentare.

saimadhav sadda
saimadhav sadda il 21 Feb 2019
please send a matlab code for logo detection???

Richiesto:

il 26 Dic 2013

Risposto:

il 21 Feb 2019

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by