Palmprint investigation has been studied by many people over10 years. During this study, many different issues related to palmprintrecognition have been found and resolved. The papers related to palmprintrecognition provides an overview of palmprint research, describing inparticular capture devices, preprocessing, verification algorithms, palmprint-related fusion, algorithms especially designed for real-time palmprintidentification in numerous databases and measures for protecting the usersprivacy related to palmprint identification and many suggestions are offered.The pictorial view of the surface of the palm normallycontains three flexion creases, secondary creases and ridges. The flexioncreases are also called principal lines and the secondary creases are calledwrinkles.
The flexion and the major secondary creases are formed between thethird and fifth months of pregnancy 1 and superficial lines appear afterbirth. Although the three major flexions are genetically dependent, but othercreases are not 2. Even identical twins have different palmprints 2. Thesenon-genetically deterministic and complex patterns are very useful in personalidentification.
Human beings were interested in palm lines for fortune tellinglong time ago and now also. Scientists found that palm lines are associatedwith some genetic diseases including Down syndrome, Aarskog syndrome, Cohensyndrome and fetal alcohol syndrome 68. Palmprint research employs eitherhigh or low resolution images. High resolution images are suitable for forensicapplications such as criminal detection 3. Low resolution images are moresuitable for civil and commercial applications such as access control.Generally, high resolution refers to 400 dpi or more and low resolution refersto 150 dpi or less.
Researchers can extract ridges, singular points and minutiapoints as essential features from high resolution images while in lowresolution images they generally extract principal lines, wrinkles and texture.Initially palmprint research focused on high-resolution images 4,5 but nowalmost all research is working on low resolution images for civil andcommercial applications. The design of a biometric system takes account of fiveobjectives: cost, user acceptance and environment constraints, accuracy, com-putation speed and security. Reducing accuracy can lead to increase of speed.Typical examples are hierarchical approaches. Reducing user acceptance can helpin improving accuracy.
For instance, users are required to provide more samplesfor training. Increasing in cost can lead to enhance in security. We can embedmore sensors to collect different signals for liveness detection. In someapplications, environmental constraints such as memory usage, powerconsumption, size of templates and size of devices have to be fulfilled. A biometricsystem installed in Personal Digital Assistant requires low power and memoryconsumption but these requirements may not be sufficient for biometric accesscontrol systems. A practical biometric system should balance all of the aboveaspects.
A typical palmprint recognition system consists of five parts:palmprint scanner, preprocessing, feature extraction, matcher and database. Thepalmprint scanner function is to collect palmprint images. Preprocessing helpsin setting up a coordinate system to align palmprint images and to segment a partof palmprint image for feature extraction. Feature extraction motive is toobtain effective features from the preprocessed palmprints. A matcher comparestwo palmprint features and a database stores registered templates.