Hello all,
I am working on Low resolution Face recognition with Occlusion handling.
My question is in two folds:
Question 1. Which of the combination of feature extraction will work well on low resolution images with occlusion: Gabor and Histogram of Oriented Gradients (HOG), or SURF and HOG features when applied to other techniques to match the features.
Question 2. I want a code to combine either of these features for better results.
Suggestions are welcomed.
Thanks
Hello dear,
ii'm worked on this field before. let me ask you some question :
1- do you extracted the feature for each method or not ???
2- in your model you have separate each one like GABOR or SURF ??
@ Mohammed Alzaidi, Thanks for you response, I appreciate.
1. Yes, I have extracted features from my Training set images
2. Yes, I have separate Feature extraction Technique like Gabor and HOG, or SURF and HOG.
I am new to Face recognition, my idea is to extract features from my training set and fuse the Extracted features using DCA after Dimensionality reduction to obtain a more discriminate information.
Should you have a better approach or guide to the problem, I will appreciate your suggestions.
Thanks
1- extract feature first for you dataset .
2- i recommended for you use LBP distributor and sift this is may be give you more feature and you can ues weka softwear this is has alot of classifier to help you do comparative for that .
hogfeatures = [1 2 3 4 ........];
surffeatures=[1 2 3 4...........];
hog_surffeatures = [hogfeatures surfeatures];
ans:
[1 2 3 4 ......................... 1 2 3 4.....];
hog_surffeatures = [hogfeatures; surfeatures];
ans:
[1 2 3 4 .........................
1 2 3 4 ......................... ]
in this way you can combine the features using MATLAB
wish you best of luck.
with regards,
SATISH
Wow Interesting! Thanks Satish Kumar Mallappa. I appreciate
- Badges
- Science topic
- Similar topics
- Analytical Chemistry
- Extraction