Browsing by Author "Rushdi, Iman Raad Rushdi"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
Item FACE RECOGNITION USING IMAGE PROCESSING AND MACHINE LEARNING METHODS(2022-01-20) Rushdi, Iman Raad Rushdi; ŞENGÜL, GökhanThe human face is a complex multidimensional visual construct, which makes it very challenging to create a computational model for recognition. Basically; face recognition is a method of recognizing a person based on the image of his or her face and has become an important area of study, covering various subjects such as image processing, computer vision and machine learning. The main challenge with facial recognition is how to correctly identify the correct feature for facial detection. This study presents an approach for the recognition of the human face based on the features extraction from the image. The face recognition system has been applied on ORL and YALE datasets. The proposed method was initially implemented in three steps. For pre-processing phase, Discrete Wavelet Transform (DWT) with Daubechies transform was applied. At second step, feature extraction phase was implemented based on Local Binary Pattern (LBP) and Gray Level Co-Occurrence Matrix (GLCM). Third step, Euclidean Distance was implemented for classification phase. Moreover, the same experiments were implemented applying Particle Swarm Optimization (PSO) for feature selection approach. The study observed several conclusions: for the first experiments; implementation of DWT and LBP, when the number of training image increased; the performance rate has been increased too, rather than implementing DWT, LBP and GLCM methods that conducted 82.50% of recognition rate when implementing on ORL database and 90% when implementing the three methods on YALE database. However, the implementation of PSO algorithm have increased the accuracy rate up to 95% for ORL database and 93% on YALE database.