High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

An Advance Subspace Method For Implementing Palm Print Recognition


Abhishek Singhal , modern institute of technology & research centre; manish mukhija, modern institute of technology & research centre


Palm Print Recognition, 2DPCA, Alternate 2DPCA, Kernel 2DPCA, (2D*2D) PCA


Biometrics are very useful to identify individuals based on their behavioural and physiological features, that can be used for their personal authorization. Various physical features like, iris patterns, retina patterns, palm print patterns, fingerprint patterns, facial features etc. are used for such purposes. Palm print identification involves recognizing an specific by matching the various wrinkles, principal lines and creases on the surface of the palm of the hand. The base for using the palm prints lies in the fact that since palm print patterns are generated by random orientations of tissues and muscles of the hand during birth, no two individuals have exactly the same palm print pattern. Research of Palm print can be possible with both low resolution and high resolution images. Low resolution images are more appropriate for commercial and civil applications such as financial transaction, access control, etc. while High resolution images are appropriate for forensic applications such as criminal detection. Generally speaking, high resolution rises to four hundred dpi or more while low resolution rises to one hundred and fifty dpi or less. Researchers can extract generally principal lines, wrinkles and texture in low resolution while from high resolution images features are extracted as ridges, singular points and minutia points. In this paper a comparative study for palm print features of different subspace methods have been projected. Where the different subspace methods are separately exploited by using a classifier-Euclidean distance to find the algorithm performance. The experiment results by using two palm print databases determine that the proposed method of class specific information with 2D-PCA, alternate 2DPCA, Kernel PCA and (2D*2D) PCA, Where in comparison to other algorithm (2D*2D) PCA method provides the better results and the recognition rate by this method is given around 87 percent.

Other Details

Paper ID: IJSRDV4I110052
Published in: Volume : 4, Issue : 11
Publication Date: 01/02/2017
Page(s): 7-11

Article Preview

Download Article