MATLAB Answers

Using Singular value decomposition for feature extraction from images

15 views (last 30 days)
Ron Herman
Ron Herman on 17 Apr 2020
Answered: Christine Tobler on 17 Apr 2020
Suppose I have a single subject who has 10 images under different conditions. I need to apply singular value decomposition on all 10 images for feature extraction. Then storing the common extracted feature from 10 images.
The purpose is to save space. Rather than saving all 10 images I would like to save common extracted in one place and use the common extracted feature for future face recognition of the same subject (person).
The 10 images have been attached.

  2 Comments

Christine Tobler
Christine Tobler on 17 Apr 2020
Do you want to compute the singular value decomposition of each of these images and store those? Or compute a combined decomposition of all 10 images?
Ron Herman
Ron Herman on 17 Apr 2020
SVD on each image for feature extraction.
I also want to know feature extraction using SVD

Sign in to comment.

Answers (1)

Christine Tobler
Christine Tobler on 17 Apr 2020
Image compression using SVD is a pretty common example (although not the most efficient way to compress an image), here are some relevant links:
For feature extraction, I don't know of a "standard" algorithm to do this using SVD. I found some papers searching for this topic, maybe one of these could be helpful:

  0 Comments

Sign in to comment.


Translated by