i have an image of a face, I want to place feature points in the image as shown below.
i've attched the images please go through.
Any help would be appreciated, thank you
Various features you can calculate: correlation coefficient, Energy, entropy etc.
BUT you need a technique for representation. you can also apply PCA directly to calculate features.
hi you can use subspace methods .
because subspace methods converts / transforms into problem space like invariant problem space and variant problem space.
Detection of face belongs to invariant space.
use PCA ,ICA , LDA
please go through the following
http://www.face-rec.org/interesting-papers/general/tr2004-041.pdf
http://www.cc.gatech.edu/~hic/8803-Fall-09/slides/SubSpace-Learning.pdf
This is a very useful feature bank. http://www.robots.ox.ac.uk/~vgg/research/texclass/filters.html
The matlab code for the feature properties is provided. A simple cross-correlation (or convolution) with the test image will give you the points of maximum response.
Hi,
You can detect the feature points based on the clarity of the image, size of the feature in an image using the simple techniques like correlation coefficient, entropy measurement, or by defining a feature some geometric parameter for its recognition. You can use principle component analysis, wavelet transform to achieve feature detection. You can follow my papers on Image processing to get an idea. Hope you can get a better idea.
https://www.researchgate.net/profile/Prasoon_Kumar3
with regards,
Prasoon
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing