The Incremental Probabilistic Neural Network

Kou, Jialiang; Xiong, Shengwu; Wan, Shuzhen; Liu, Hongbing

doi:10.1109/icnc.2010.5583589

Cited by 10 publications

(6 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Within the first hidden layer of the PNN, the node outputs are calculated as [24, 25]

Z_{i, j} = \exp (][, - \frac{{false(x - {bold-italicw}_{i, j} false)}^{normalT} false(x - {bold-italicw}_{i, j} false)}{2 σ^{2}}), i = 1, 2, \dots, p, j = 1, 2, \dots, c

where

Z_{i, j}

is the probability output value of a hidden node, x is the input vector

x = false[x_{1}, x_{2}, \dots, x_{n} {false]}^{normalT}

{bold-italicw}_{i, j}

is the i th element from class j , that is, the training vector

{bold-italicw}_{i, j} = false[w_{1}, w_{2}, \dots, w_{n} {false]}^{normalT}

, p represents the number of input training patterns for each class and c is the number of classes.…”

Section: Pnn and Missing Featuresmentioning

confidence: 99%

Robust feature extraction and salvage schemes for finger texture based biometrics

Al-Nima¹,

Dlay

Al-Sumaidaee

et al. 2016

IET biom.

View full text Add to dashboard Cite

In this paper, an efficient human authentication method is proposed which utilizes Finger Texture (FT) patterns. This method consists of two essential contributions: a robust and automatic finger extraction method to isolate the fingers from the hand images; and a new feature extraction method based on an Enhanced Local Line Binary Pattern (ELLBP). To overcome poorly imaged regions of the FTs a method is suggested to salvage missing feature elements by exploiting the information embedded within the trained Probabilistic Neural Network (PNN) used to perform classification. Three databases have been applied in this paper: PolyU3D2D, IIT Delhi and spectral 460 from Multi-spectral CASIA images. Experimental studies show that the best result was achieved by using ELLBP feature extraction. Furthermore, the salvaging approach proved effective in increasing the verification rate.

show abstract

“…Within the first hidden layer of the PNN, the node outputs are calculated as [24, 25]

Z_{i, j} = \exp (][, - \frac{{false(x - {bold-italicw}_{i, j} false)}^{normalT} false(x - {bold-italicw}_{i, j} false)}{2 σ^{2}}), i = 1, 2, \dots, p, j = 1, 2, \dots, c

where

Z_{i, j}

is the probability output value of a hidden node, x is the input vector

x = false[x_{1}, x_{2}, \dots, x_{n} {false]}^{normalT}

{bold-italicw}_{i, j}

is the i th element from class j , that is, the training vector

{bold-italicw}_{i, j} = false[w_{1}, w_{2}, \dots, w_{n} {false]}^{normalT}

, p represents the number of input training patterns for each class and c is the number of classes.…”

Section: Pnn and Missing Featuresmentioning

confidence: 99%

Robust feature extraction and salvage schemes for finger texture based biometrics

Al-Nima¹,

Dlay

Al-Sumaidaee

et al. 2016

IET biom.

View full text Add to dashboard Cite

show abstract

“…4. Principally, the calculations of the APNN are as follows: After constructing the input values, hidden or pattern layer values are extracted according to [35,36]:…”

Section: Roi Of Middlementioning

confidence: 99%

Finger Texture Verification Systems Based on multiple spectrum Lighting Sensors with Four Fusion Levels

Al-Kaltakchi

Omar

Abdullah

et al. 2019

Iraqi Journal of ICT

View full text Add to dashboard Cite

Finger Texture (FT) is one of the most recent attractive biometric characteristic. Itrefers to a finger skin area which is restricted between the fingerprint and the palm print (just after including the lower knuckle). Different specifications for the FT can be obtained by employing multiple images spectrum of lights. This inspired the insight of applying a combination between the FT features that acquired by utilizing two various spectrum lightings in order to attain high personal recognitions. Four types of fusion will be listed and explained here: Sensor Level Fusion (SLF), Feature Level Fusion (FLF), Score Level Fusion (ScLF) and Decision Level Fusion (DLF). Each fusion method is employed and examined for an FT verification system. From the database of Multiple Spectrum CASIA (MSCASIA), FT images have been collected. Two types of spectrum lights have been exploited (the wavelength of 460 nm, which represents a Blue (BLU) light, and the White (WHT) light). Supporting comparisons were performed, including the state-of-the-art. Best recognition performance were recorded for the FLF based concatenation rule by improving the Equal Error Rate (EER) percentages from 5% for the BLU and 7% for the WHT to 2%.

show abstract

“…The RPNN operations can be explained as follows: the hidden values of the hidden layer are calculated according to (7) [26], this equation is also known as the RBF

\begin{matrix} right leftthickmathspace.5em \\ Z_{i, j} & = normalexp [- \frac{{(x^{λ} - w_{i, j}^{λ})}^{T} (x^{λ} - w_{i, j}^{λ})}{2 σ^{2}}], \\ i = 1, 2, \dots, p, j = 1, 2, \dots, c \end{matrix}

where

Z_{i, j}

represents a hidden layer node value,

{bold-italicx}^{λ}

represents the input vector

{bold-italicx}^{λ} = false[x_{1}^{λ}, x_{2}^{λ}, \dots, x_{n}^{λ} false]^{normalT}

{bold-italicw}_{i, j}^{λ}

is the

i th

vector of class j ; thus, the weight vector can be described or formed as

{bold-italicw}_{i, j}^{λ} = false[w_{1}^{λ}, w_{2}^{λ}, \dots, w_{n}^{λ} false]^{normalT}

λ

represents the certain spe...…”

Section: Re‐enforced Pnnmentioning

confidence: 99%

Personal verificationbased on multi‐spectral finger texture lighting images

Al-Nima¹,

Al-Kaltakchi

Al-Sumaidaee

et al. 2018

IET signal process.

View full text Add to dashboard Cite

Finger texture (FT) images acquired from different spectral lighting sensors reveal various features. This inspires the idea of establishing a recognition model between FT features collected using two different spectral lighting forms to provide high recognition performance. This can be implemented by establishing an efficient feature extraction and effective classifier, which can be applied to different FT patterns. So, an effective feature extraction method called the surrounded patterns code (SPC) is adopted. This method can collect the surrounded patterns around the main FT features. It is believed that these patterns are robust and valuable. Furthermore, a novel classifier termed the re-enforced probabilistic neural network (RPNN) is proposed. It enhances the capability of the standard PNN and provides better recognition performance. Two types of FT images from the multi-spectral Chinese Academy of Sciences Institute of Automation (CASIA) database were employed as two types of spectral sensors were used in the acquiring device: the white (WHT) light and spectral 460 nm of blue (BLU) light. Supporting comparisons were performed, analysed and discussed. The best results were recorded for the SPC by enhancing the equal error rates at 4% for spectral BLU and 2% for spectral WHT. These percentages have been reduced to 0% after utilising the RPNN.

show abstract

The Incremental Probabilistic Neural Network

Abstract: With

Cited by 10 publications

References 5 publications

Robust feature extraction and salvage schemes for finger texture based biometrics

Robust feature extraction and salvage schemes for finger texture based biometrics

Finger Texture Verification Systems Based on multiple spectrum Lighting Sensors with Four Fusion Levels

Personal verificationbased on multi‐spectral finger texture lighting images

Contact Info

Product

Resources

About