Dense facial landmark detection is a key element of face processing pipeline. In this paper we survey and analyze modern neural-network-based facial landmark detection algorithms, focus on approaches that have led to a significant increase in quality over the past few years on datasets with large pose and emotion variability, significant face occlusion, all of which are typical in real-world scenarios. We summarize the improvements into categories, provide quality comparison on difficult and modern in-the-wild datasets: 300 Faces in-the-wild (300W), Annotated Facial Landmarks in-the-wild (AFLW), Wider Facial Landmarks in-the-wild (WFLW), Caltech Occluded Faces in-the-wild (COFW). Additionally, we compare algorithm speed on desktop central and graphical processors, mobile devices. For completeness, we also briefly touch on established methods with open implementations available. Besides, we cover applications and vulnerabilities of the landmark detection algorithms. We hope that generalizations that we make will lead to further algorithm improvements.

Context. Neural networks require a large amount of annotated data to learn. Meta-learning algorithms propose a way to decrease number of training samples to only a few. One of the most prominent optimization-based meta-learning algorithms is MAML. However, its adaptation to new tasks is quite slow. The object of study is the process of meta-learning and adaptation phase as defined by the MAML algorithm.Objective. The goal of this work is creation of an approach, which should make it possible to: 1) increase the execution speed of MAML adaptation phase; 2) improve MAML accuracy in certain cases. The testing results will be shown on a publicly available few-shot learning dataset CIFAR-FS.Method. In this work an improvement to MAML meta-learning algorithm is proposed. Meta-learning procedure is defined in terms of tasks. In case of image classification problem, each task is to try to learn to classify images of new classes given only a few training examples. MAML defines 2 stages for the learning procedure: 1) adaptation to the new task; 2) meta-weights update. The whole training procedure requires Hessian computation, which makes the method computationally expensive. After being trained, the network will typically be used for adaptation to new tasks and the subsequent prediction on them. Thus, improving adaptation time is an important problem, which we focus on in this work. We introduce lambda pattern by which we restrict which weight we update in the network during the adaptation phase. This approach allows us to skip certain gradient computations. The pattern is selected given an allowed quality degradation threshold parameter. Among the pattern that fit the criteria, the fastest pattern is then selected. However, as it is discussed later, quality improvement is also possible is certain cases by a careful pattern selection.Results. The MAML algorithm with lambda pattern adaptation has been implemented, trained and tested on the open CIFAR-FS dataset. This makes our results easily reproducible.Conclusions. The experiments conducted have shown that via lambda adaptation pattern selection, it is possible to significantly improve the MAML method in the following areas: adaptation time has been decreased by a factor of 3 with minimal accuracy loss. Interestingly, accuracy for one-step adaptation has been substantially improved by using lambda patterns as well. Prospects for further research are to investigate a way of a more robust automatic pattern selection scheme.

RFID tags see a widespread use in modern security systems, including home intercoms, access control cards, contactless credit cards, biometric passports. Here we focus on a single application, namely access control systems. Currently they have either high cost or low security guarantees. Hence, the developments focusing on improving access control security while lowering the cost is a rapidly developing field. The purpose of this work is to create an alternative access control scheme, where card scanners are replaced with passive RFID tags, and all of the communication is done via user's smartphone Wi-Fi. Based on the analysis of existing approaches to the development of access control systems, it was concluded that use of mobile systems is the most promising due to their expandability and presence of a large number of sensors, such as NFC, camera etc. In the proposed model RFID tags are mounted near a turnstile or a smart door. Tag reading and programming is done via NFC chip directly on an Android or iOS mobile device, which allows for a significant price cut for such a system implementation. A detailed description of a tag writing procedure with the data required to perform it is provided. To enhance security, together with smartphone-based authorization we require the user to provide his photograph while entering a secure gate. The photograph is then displayed on a monitoring dashboard side-by-side with his registration picture, so that the two can then be matched

In many practical cases face detection on smartphones or other highly portable devices is a necessity. Applications include mobile face access control systems, driver status tracking, emotion recognition, etc. Mobile devices have limited processing power and should have long-enough battery life even with face detection application running. Thus, striking the right balance between algorithm quality and complexity is crucial. In this work we adapt 5 algorithms to mobile. These algorithms are based on handcrafted or neural-network-based features and include: Viola-Jones (Haar cascade), LBP, HOG, MTCNN, BlazeFace. We analyze inference time of these algorithms on different devices with different input image resolutions. We provide guidance, which algorithms are the best fit for mobile face access control systems and potentially other mobile applications. Interestingly, we note that cascaded algorithms perform faster on scenes without faces, while BlazeFace is slower on empty scenes. Exploiting this behavior might be useful in practice.

No abstract