Make biometric mobile applicable to traditional indian palmistry

• Main Author: Monica Cheryl Sandra Premkumar.
• Other Authors: Li Fang
• Format: Final Year Project
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: http://hdl.handle.net/10356/52570
• Institution: Nanyang Technological University

Biometrics is a study done through statistical analysis of physical and behavioural characteristics allowing us to identify a person. There are several biometric identification systems based on physical attributes such as fingerprint, hand geometry, palmprint and analysis of capillary blood vessels in the retina. Amongst these, palmprint has the highest potential to be more significant and applicable in the real world. While, analysis of palmprints paves way to create a sound authentication system, the analysis of palmprints can also be applied to traditional Indian palmistry. Indian palmistry is a practice that has been preserved for thousands of years and still bears significance in today’s context in several eastern cultures and garners interest from the west as well. The main focus of this project is to review the existing palmprint/hand vein Biometric Identification System and evaluate the applicability of Indian Palmistry. This will be done by assessing the accuracy of image matching in the existing Biometric Identification System through a few experiments using the image databases that are already available. These databases include palmprint images of good quality, lower quality from mobile phones and distorted images. The experiments will be conducted with images of different quality to assess the system’s adaptability to those images and its sensitivity to the algorithm. After assessing the accuracy rates, the next main objective is to analyse the means to integrate the palmprint identification system into the Fortune Telling Application. This Fortune Telling Application is very similar to the Biometric Identification System as they both use the same concept of matching palmprint patterns. While the processes in both systems are similar in concept, the line extraction module however differs. In the Biometric Identification System, feature extraction involves 6 steps where it is able to extract as many edges as possible and condensed into a line edge map (LEM). However that process does not apply to feature extraction in the Fortune Telling Application since only the principal lines need to be extracted. Therefore Hough transformation is used where the straight line segments are detected and the lines with similar parameters are grouped to form the principal lines. With the principal lines extracted, the Fortune Telling Application can successfully match the pattern with the closest model in the database and return a prediction. In this project, the Biometric Identification System has been successfully integrated into the Fortune Telling Application and has a lot of potential for future use and application. The next step in expanding this project would be to increase the model database in this system and evaluate its performance quantitatively. Following that, other parameters of palmistry such as age and palm regions could be studied and implemented in the future. Inevitably, with a sound system in place, a mobile application could be developed to read palmprints and provide instantaneous predictions and insights.