Computational Intelligence and Neuroscience

Research Article

A Gabor-Block-Based Kernel Discriminative Common Vector Approach Using Cosine Kernels for Human Face Recognition

Table 4

Specificity and sensitivity measure of the FERET dataset.


Total no. of classes = 200, Total no. of images = 3600
		Individual belonging to a particular class

		Using first 3 images of an individual as training images
		Positive	Negative

	Positive
FERET test	Negative


		Using first 4 images of an individual as training images
		Positive	Negative

	Positive
FERET test	Negative

So considering the first 4 images in Figures 4(a)–4(d) of a particular individual for training the achieved rates are:
False positive rate = /( + ) = 1 − Specificity = .72%.
False negative rate = /( + ) = 1 – Sensitivity = 6%.
Accuracy = ( + )/( + + + ) ≈ 96.6%.
So considering in the first 3 images Figures 4(a)–4(c) of a particular individual for training the achieved rates are:
False positive rate = /( + ) = 1 − Specificity = 1.2%.
False negative rate = /( + ) = 1 – Sensitivity = 6.75%.
Accuracy = ( + )/( + + + ) ≈ 96.1%.