Improved crossover firefly algorithm based deep Beleif network for low-resolution face recognition
- PDF / 1,183,963 Bytes
- 21 Pages / 439.37 x 666.142 pts Page_size
- 103 Downloads / 180 Views
Improved crossover firefly algorithm based deep Beleif network for low-resolution face recognition Wael Mohammad Alenazy 1 & Abdullah Saleh Alqahtani 1 Received: 23 March 2020 / Revised: 8 August 2020 / Accepted: 24 September 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
Face detection by low-resolution image (LR) is one of the key aspects of HumanComputer Interaction(HCI). Due to the LR image, which has changes in pose, lighting, and illumination, the performance of face recognition is reduced. In this work, we propose the Deep Belief Network-Crossover based Firefly (DBN-CROFF) method for face recognition from low-resolution images. The Histogram of Gradient (HOG) and 2Dimensional Discrete Wavelet Transform (2D-DWT) to extract facial width, size of the cheeks, skin tone, nose, and lip shape features from facial data. The Kernel Principle Component Analysis (k-PCA) is used to successfully reduce the dimension of the feature. The experimental performance of the proposed method is evaluated using four datasets namely LFW, Multi-PIE, Extended Yale-B, and FERET with conventional techniques. Finally, the proposed DBN-CROFF solution surpasses the other conventional facial recognition approaches by giving a higher accuracy of recognization. Keywords Low-resolution image . Face recognition . Facial features . DBN . CROFF
1 Introduction One of the important research fields in computer vision is face recognition that made many valuable contributions in recent years. Face recognition systems are used to recognize a person from a mug, video, shot, and digital images. Recently, most commonly and popularly used biometric technology is face image recognition. Generally, the face recognition model mainly concentrates on large sufficient frontal faces that consist of enough data for recognition [11]. The attributes of face recognition technologies are significantly higher reliability and low invasiveness of the acquisition. One of the major challenges faced by the facial recognition systems is the identification of faces from the LR images [40] captured from a longer distance. The discriminate face properties are tainted due to the low-resolution image that considerably reduces the traditional face recognition accuracy. * Wael Mohammad Alenazy [email protected]; [email protected]
1
Department of Self Development Skills, CFY Deanship King Saud University, Riyadh, Saudi Arabia
Multimedia Tools and Applications
Practically, the face recognition model deals with poor image quality or low resolution (LR) images and also larger variations in lighting conditions, facial expressions, and pose. Hence, the face recognition using low resolution is a challenging task. The conservative face recognition models are working well for high-quality images under suitable environmental conditions [1]. By using deep learning methods, the high-resolution face images produce over 99% of recognition rates. Nevertheless, face images including limited information, which for LowResolution Face Recognition(L
Data Loading...
