Face Recognition and Detection through Similarity Measurements

Abstract

The facial recognition has been a problem worked on around the world for many persons, the problem has emerged in multiple fields and sciences, especially in computer science and other fields that are very interested in this technology are robotic, criminalist etc. Unfortunately, many reported face recognition techniques relay on the size and representative of training set such as e-passport, law enforcement and id-card identification, and most of them will suffer serious performance drop if only one training sample per person is available to the systems [1].In a face image, only a part of face is changed due to pose, illumination and other source of changes .In this paper, a novel face recognition detection approach known as Gabor wavelet based PCA approach is presented based on fusing global and local features of image. To extract global and local features, Gabor wavelet filter are applied on the whole image and non-overlapping sub-images with equal size. To reduce the dimension of new fused feature vector and to better characterize the similarity between each gallery face and the probe image set, Principal Component Analysis (PCA) is employed. And finally, measure the similarity between the images by using the Euclidean distance as classifier. The Experimental results shows that proposed technique improves the efficiency of face recognition under varying illumination, expression and variation in poses of face images by using standard databases when compared to traditional PCA and Conventional method such as global Gabor faces recognition. In this paper, the proposed algorithm is tested on the public and largely used ORL database.

Enhancing Building Energy Management: Adaptive Edge Computing for Optimized Efficiency and Inhabitant Comfort

“…Thanks to deep neural networks [93], new milestones have been achieved both in the prediction [94] and classification of data [95] (e.g., generative adversarial networks (GANs) [96]) as well as in the processing of long series of data (e.g., recurrent networks [97] and long short-term memory (LSTM) [98]). These cuttingedge solutions [99] (including auto-encoders, deep belief networks, deep forest, capsules, deep Boltzmann machines, and hybrid algorithms merging some of these solutions) have high success rates in pattern recognition [100], artificial vision [101], natural language processing (NLP) [102], data relations, etc. The superposition of different layers for the analysis of big datasets allows such algorithms to cluster [103] and recognize (low level) features in (high level) data (audio, text, images, etc.…”