An efficient automatic palm reading algorithm and its mobile applications development

Palm reading is a traditional practice in China for a few thousand years to tell ones' fortune. Currently, there is a lack of mobile applications that allow palm reading to be done automatically and efficiently. This study aimed at developing an effective palm reading algorithm which can run in an Android platform efficiently. OpenCV and Java were used for the implementation. Our palm reading algorithm uses an adaptive thresholding approach to segment the palm image from the background, extract the fingers and calculate their length, extract the three principal palm lines in which regression is applied to produce connected and continuous palm lines. The algorithm was implemented as an Android application. Results showed that the algorithm can be run within 2 to 4 seconds, and the automatic palm reading can be done on mobile platforms accurately. The study enriched existing market of mobile applications that aim at palm reading. With successful implementation of such platform, and by collecting more personal information of the users, such as personality and health status, this application can be applied for future research on the prediction of personality and health.