Machete

Face recognition with single training sample has the problem that the sample is single and face pattern is changeable. The rate of traditional face recognition methods declines seriously in the case of single training sample. Thus, studying the face recognition method which is appropriate for single training sample is challenging in face recognition field. In this paper, in order to solve the problem of single training sample, sample extension method is adopted to expand fourteen virtual samples. And for the problem of changeable face pattern, like pose, expressions, illumination etc., this paper uses image blocking and applies weighted two-directional 2DPCA to extract the local feature information of face image. Further, fuse the local feature information of sub-image to face recognition. The experimental results on ORL and YALE face database show that the method in this paper achieves good recognition effect.

Due to its wide applications in practice, face recognition has been an active research topic. With the availability of adequate training samples, many machine learning methods could yield high face recognition accuracy. However, under the circumstance of inadequate training samples, especially the extreme case of having only a single training sample, face recognition becomes challenging. How to deal with conflicting concerns of the small sample size and high dimensionality in one-sample face recognition is critical for its achievable recognition accuracy and feasibility in practice. Being different from the conventional methods for global face recognition based on generalization ability promotion and local face recognition depending on image segmentation, a single-sample face recognition algorithm based on locality preserving projection (LPP) feature transfer is proposed here. First, transfer sources are screened to obtain the selective sample source using the whitened cosine similarity metric. Second, we project the vectors of source faces and target faces into feature subspace by LPP, respectively, and calculate the feature transfer matrix to approximate the mapping relationship on source faces and target faces in subspace. Then, the feature transfer matrix is used on training samples to transfer the original macro characteristics to target macro characteristics. Finally, the nearest neighbor classifier is used for face recognition. Our results based on popular databases FERET, ORL, and Yale demonstrate the superiority of the proposed LPP feature transfer-based one-sample face recognition algorithm when compared with popular single-sample face recognition algorithms, such as (PC) $^{2}\text{A}$ and Block FLDA.

Attentional selectivity focuses on what is currently relevant. Relevance changes frequently in everyday life, triggering rapid reassignments of attentional priorities. Such reassignments are often not associated with behavioural changes and are thus difficult to assess objectively. Here, we measured rapid, covert switches between preparatory task settings (attentional templates) in visual search, as they occurred in real time. Participants searched for colour-defined targets in search displays that appeared unpredictably either early (after 700 ms) or late (after 1500 ms) on each trial. In Experiment 1, early and late targets were defined by different colours. Participants first had to activate a template for the early target colour, then switch to a template for the late target colour if no early search display appeared. In Experiment 2, cues signalled whether the initial target template had to be maintained or changed. Template activation states were tracked with N2pc components to rapid sequences of irrelevant probes matching either the early or late target colour. A template for the early target colour was active from about 300 ms before the expected arrival of early search displays, followed by a template switch. Switches based on endogenous temporal expectations emerged more gradually in time than switches in response to external cues. Presenting cues in Experiment 2 triggered a temporary search template deactivation even when the target colour remained unchanged, indicating that template maintenance is subject to an attentional blink. Results demonstrate that rapid switches between attentional templates in visual search can be tracked with high temporal precision.

Similarity mapping for robust face recognition via a single training sample per person

In this paper, an adaptive approximation image reconstruction method based on orthogonal triangular with column pivoting (QRCP) decomposition algorithm is proposed for single sample problem in face recognition. By using QRCP the single sample and its transpose are decomposed to two sets of basis images. Then an adaptive approximation image reconstruction method is proposed to reconstruct two approximation images from the two basis image sets respectively. The single training sample and its two approximation images of each object form a new training set, which can make the fisher linear discriminant analysis (FLDA) be applied to single sample problem in face recognition. The performance of the proposed method is verified on Yale, FERET, and ORL face databases. The experimental results indicate that the proposed method is efficient and outperforms some existing methods which are proposed to overcome the single sample problem.

Non-linear dictionary representation of deep features for face recognition from a single sample per person

Surveillance video face recognition with single sample per person based on 3D modeling and blurring

Cycle-autoencoder based block-sparse joint representation for single sample face recognition

For the single training sample per person (SSPP) problem, this paper proposes an adaptive weighted LTP algorithm with a novel weighted method involving the standard deviation of the sub-images’ feature histogram. First, LTP operator is used to extract texture feature and then feature images are split into sub images. Then, standard deviation is used to compute the adaptive weighted fusion of features. Finally, the nearest classifier is adopted for recognition. The experiments on the ORL and Yale face databases demonstrate the effectiveness of the proposed method.

Lack of training samples is the primary cause of low accuracy in the face recognition problem. Especially when only one labeled sample per person is available, many algorithms suffer a slump on performance or even fail to work. In order to increase the accuracy in single training samples recognition problem, virtual samples are generated to enlarge the training set. In this paper, we generate the virtual sample by subtracting the test sample from each training sample and proposed the improved CRC for single training sample face recognition. The experiment results on ORL and FERET face database are provided to validate the effectiveness and robustness of the proposed method.

Background. Ultra-Reliable Low Latency Communication (URLLC) as a service offered by fifth and sixth generation (5G/6G) wireless systems is a technological response to the needs of various mission-critical applications that require reliable data transmission with low latency. These applications also include Intelligent Transportation Systems services, which, among other things, provide connectivity and autonomous vehicle control. The combination of high reliability and low latency requirements in URLLC usage scenarios creates a security problem for URLLC data transmission that cannot be solved using conventional complex cryptographic methods based on a secret key. The article discusses in detail the approach to using physical layer security mechanisms (physical layer security - PLS) as a powerful alternative to classical cryptographic security methods for URLLC, and also proposes the application of the wire-tap channel concept in URLLC with an analysis of the efficiency that can be achieved for physical layer security. Objective. The aim of the article is to provide an overview of information security solutions in URLLC usage scenarios, as well as to propose a constructive method for information protection for reliable data transmission with low latency without the use of cryptographic mechanisms based on a secret key. Methods. Theoretical research in the field of the branch channel concept was used to create solutions that allow data protection with information-theoretic stability in URLLC usage scenarios for providing IoT, connected car and autonomous driving services. Results. The article examines in detail the data security issues in ultra-reliable low latency communication (URLLC) as a service offered by fifth and sixth generation (5G/6G) wireless systems. It is determined that URLLC is a technological response to the needs of various critical applications that require reliable signal transmission with low latency, and among these applications are Intelligent Transportation Systems services, which, among other things, provide connectivity and autonomous vehicle control. It is shown that the combination of high reliability and low latency requirements in URLLC scenarios creates a security problem for URLLC data transmission that cannot be solved using conventional complex cryptographic methods based on a secret key. The feasibility of using physical layer security mechanisms (PLS) as a powerful alternative to classical cryptographic security methods for URLLC is substantiated. The approach to applying the concept of a wire-tap channel in URLLC is considered in detail, as well as the results that can be achieved for physical layer security, and the influence of code parameters for probabilistic cryptographic transformations in accordance with the concept of a wire-tap channel on PLS URLLC. Estimates of the effectiveness of PLS URLLC for finite block length codes are provided. Conclusions. An effective way to ensure data security for ultra-reliable low-latency physical layer link (PLS URLLC) of fifth-generation 5G wireless systems in the field of connected cars and vehicles of 4-5 levels of automation can be approaches based on the concept of a tapped channel ("wire-tap channel").

The computational understanding of continuous human movement plays a significant role in diverse emergent applications in areas ranging from human computer interaction to physical and neuro-rehabilitation. Non-visual feedback can aid the continuous motion control tasks that such applications frequently entail. An architecture is introduced for enabling interaction with a system that furnishes a number of gestural affordances with assistive feedback. The approach combines machine learning techniques for understanding a user's gestures with a method for the display of salient features of the underlying inference process in real time. Methods used include a particle filter to track multiple hypotheses about a user's input as the latter is unfolding, together with models of the nonlinear dynamics intrinsic to the movements of interest. Non-visual feedback in this system is based on a presentation of error features derived from an estimate of the sampled time varying probability that the user's gesture corresponds to the various tracked state trajectories in the different dynamical systems. We describe applications to interactive systems for human gait analysis and rehabilitation, a domain of considerable current interest in the movement sciences and health care.

A face recognition method by using large and representative datasets is presented in this paper. The importance of research on face recognition is fueled by both its scientific challenges and its potential applications. In this contribution, we proposes several approaches to deal with some of the difficulties that one encounters when trying to recognize frontal faces in unconstrained domains and when only one sample per class is available to the learning system. It is possible for an automatic recognition system to compensate for imprecisely localized, partially expression variant faces even when only one single training sample per class is available. Finally, we have shown that the results of an appearance-based approach totally depend on the differences that exist between the facial expressions displayed on the learning and testing images.

In this article, we propose a multifeature learning method to jointly learn compact multifeature codes (LCMFCs) for palmprint recognition with a single training sample per palm. Unlike most existing hand-crafted methods that extract single-type features from raw pixels, we first form the multi-type data vectors such as the direction-data, and texture-data to completely sample the multiple information of a palmprint image. Then, we learn the discriminative multifeatures from multi-type data vectors by maximizing the inter-palm distance, and minimizing the energy loss between the learned codes, and the original data. Moreover, our LCMFC method adaptively learns the optimal weights of multi-type features to jointly learn the compact multifeature codes. Finally, we cluster the nonoverlapping blockwise histograms of the compact multifeature codes into a feature vector for palmprint representation. Extensive experimental results on six benchmark palmprint databases are presented to show the effectiveness of the proposed method.

Rapid growth of social networks has provided an extraordinary medium to share a large volume of photographs online. This calls for designing efficient face recognition techniques that are applicable to images with low resolutions and arbitrary poses. This paper proposes a new pose invariant face recognition method for low resolution images using only a single training sample. A 3D model, reconstructed using Generic Elastic Model (3D GEM) from a frontal view training sample, is used to generate a set of nonfrontal gallery face images. The face region of the nonfrontal query sample is then extracted using the same landmark detection technique as in the 3D GEM algorithm. Afterwards, a novel texture representation technique called Local Comparative Decimal Pattern (LCDP) is proposed to extract features from each of the training and query samples. A set of experimental results on the ORL, Georgia Tech (GT), and LFW face databases demonstrates the efficiency of the proposed method compared to other state-of-the-art approaches.