Clear Filters
Clear Filters

what should we do after feature extraction?

6 views (last 30 days)
sara
sara on 11 Feb 2015
Answered: Image Analyst on 11 Feb 2015
hi
I want to use SVM...I extract mean2 and variance2 from my 256*256 and now I have 2 matrix with 86*86...my feature are mean2 and variance2 ...but I dont know how can I use these feature as input of my SVM...I search about SVM and in these example inputs are like vector...Can you guide me to use these feature as inputs??I have about 23 images..
  2 Comments
Image Analyst
Image Analyst on 11 Feb 2015
What do the two 86-by-86 matrices represent? So for each of the 23 images will you have a pair of these 86-by-86 matrices?
Don't use mean2 as the name of a variable if you have the Image Processing Toolbox because that is the name of a built-in function.
sara
sara on 11 Feb 2015
thank you for helpful guidance
first matrix shows mean of image that divided to 3*3 block(input block is 3*3 and output block is 1*1)...and second matrix shows variance of image that divided to 3*3 block...yes for 23 image I have pair of these matrices..I know that these feature are my inputs but I dont know how to use theme..should I do any process on these and then use them?

Sign in to comment.

Answers (1)

Image Analyst
Image Analyst on 11 Feb 2015
You're probably making your window size too small. What you're describing is one way that people do CBIR (Content Based Image Retrieval). Basically you get the mean and standard deviation of your image divided into tiles, then use them as feature vectors to find other images in a database that have a similar pattern. For example you could use it to find all pictures of a beach and sky similar to your sample image, but would not retrieve other images that had the same colors but laid out in a different pattern. Like it would get pictures with brown bottoms and blue tops but would not bring back a picture of a man in brown pants with a blue shirt. The example I saw divided the image up into 10 by 10 tiles and computed mean and stdev for hue, saturation, and value channels. So there were vectors of 6 hundred elements. You almost certainly don't need to have windows that are as tiny as 3 by 3. I'd say 100 tiles (10-by-10) is plenty. So then your feature vector is just a linear list of the means. Same for stdev. Just plug those vectors into the SVM machine. I don't have the Statistics Toolbox so I can't help anymore.

Community Treasure Hunt

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

Start Hunting!