Harmonic Rule for Measuring the Facial Similarities among Relatives
Ravi Kumar Yelwala Basavaraju, C.N. Ravi Kumar · Transactions on Machine Learning and Artificial Intelligence · 2018
The harmonic rule is a strategy used to measure the harmonic distance between any pair of images. The images have been trained and tested on individual pair of facial images respectively. The facial images have been trained in ten folds with equal number of facial images, where each fold consists of some equal number of images of both KinfaceW-I and KinfaceW-II, which is a benchmark dataset. The facial similarities among relatives are measured by employing a harmonic distance metric using K-nearest neighbors over a dataset KinfaceW. The proposed Harmonic Rule for Measuring the Facial Similarities among Relatives is a method used to determine the percentage of facial similarity between father-son, father-daughter, mother-son and mother-daughter. Also the proposed Extended Harmonic Rule for Measuring the Facial Similarities among relatives is another strategy used to determine the percentage of face similarities between son-father-grand father, son-father-grandmother, daughter- father-grandfather, and daughter-father-grandmother. The result of the proposed approach is better than other approaches like NRML and MNRML, which are contemporary works published.