Gender recognition from frontal facial images within project FP7-ICT-247525-HUMAVIPS

Comparing classification time

<< zpět

Overview

Aim of this experiment is to compare classification time of different kinds of descriptors. Exactly: LBP (Local Binary Patterns), HOG (Histogram Of Gradients) and finally HLBP (Histogram of LBP).

Results

Parameters of experiment
imSize [h x w]80 x 64
winSize [h x w]60 x 40
win positionhalf in either axis
scales0.8, 1, 1.2
angles-pi/30, -pi/60, 0, pi/60, pi/30
aspect ratios (H/W)1.5

DescriptorClassification time per imageDescriptor parameters
LBP222 msnPyramids: 4
HOG796 mscellSize: 2px, blockSize: 4, nBins = 9
HLBP348 mscellSize: 5px, overlaping: false

Resources

[1] Corinna Cortes, Vladimir Vapnik: Support-vector networks - http://www.springerlink.com/content/k238jx04hm87j80g/
[2] Vojtech Franc, Soeren Sonnenburg: LIBOCAS - Library implementing OCAS solver for training linear SVM classifiers from large-scale data - http://cmp.felk.cvut.cz/~xfrancv/ocas/html/
<< zpět