Performance Evaluation of the Eigenface Algorithm on Plain-Feature Images in Comparison with Those of Distinct Features
A A Alabi, L A Akanbi, Asma`a Al Ibrahim · American Journal of Signal Processing · 2015
This research work centered on the analysis and identification of the major features making up the human face in relation to their roles in the Eigenface Algorithm. The area of concern was to ascertain the workability and efficiency of the developed algorithm by evaluating its performance on gallery of faces with plain features in comparison with that of faces with distinct features. Seventy five percent (75%) representing three out of every four images were used to form the training set while the remaining twenty five (25%) were meant for the test images. The characteristics of the face in terms of facial dimension, types of marks, structure of facial components such as the eye, mouth, chin etc. were analyzed for identification. The face images were resized for proper reshaping and cropped to adjust their backgrounds using the Microsoft Office Picture Manager. The system code was developed and run on both set of face images (Plain and Distinct) using Matrix Laboratory software (MatLab7.0). The system was observed to be of better results with the use of faces with distinct features than those with plain features. This was duly observed both in terms of the total number of identified images as well as the execution time. Nearly all the tested images were identified from those with distinct features while the case was not the same with those with plain images. The system evaluation has shown an estimated difference of 25% in terms of identification and 45% in execution time. The study concluded that the existence of distinct features on those facial images employed catalyzed the recognition rate of the developed PCA code on such faces not only in terms of identification but also in the speed of the system. It has also shown that the performance of the Eigenface algorithm is greater in the recognition of faces with distinct features compared with those with plain features.