Personality Identification of Palmprint Using Convolutional Neural Networks
Andri Ariyanto, Esmeralda C. Djamal, Ridwan Ilyas · 2018
Overview of personality is needed such employee recruitment, counseling in schools, and the development of the children - for the parents. Palmistry is a technique through the palmprint of the hand that contains personality. Unfortunately, this way is not easy. It requires special expertise in the science of Palmistry. Deep learning one of the developments in machine learning that has a high level of recognition. This research proposed personality identification of palmprint using Convolutional Neural Networks (CNNs). The methods began with learning before being used for identification. Learning used data which were obtained from 11,000 hands that have been labeled customized according to Palmistry science, i.e., Air, Earth, Fire, and Water. The results showed that the accuracy reached 96.7% of new data and 99.7% of training data. The best accuracy was given in five convolution configurations, five Max Pooling, a hidden layer with 1024 hidden neurons, and 512 input vectors. The time required in the identification process is fast enough that it can be used properly on mobile devices for ease of use of the system.