Active and Assisted Living (AAL) technologies aim to enhance the quality of life of older adults and promote successful aging. While video-based AAL solutions offer rich capabilities for better healthcare management in older age, they pose significant privacy risks. To mitigate the risks, we developed a video-based monitoring system that incorporates different privacy-preserving filters. We deployed the system in one assistive technology center and conducted a qualitative study with older adults and other stakeholders involved in care provision. Our study demonstrates diverse users’ perceptions and experiences with video-monitoring technology and offers valuable insights for the system’s further development. The findings unpack the privacy-versus-safety trade-off inherent in video-based technologies and discuss how the privacy-preserving mechanisms within the system mitigate privacy-related concerns. The study also identifies varying stakeholder perspectives towards the system in general and highlights potential avenues for developing video-based monitoring technologies in the AAL context.

Abstract Human skin segmentation is a crucial task in computer vision and biometric systems, yet it poses several challenges such as variability in skin colour, pose, and illumination. This paper presents a robust data-driven skin segmentation method for a single image that addresses these challenges through the integration of contextual information and efficient network design. In addition to robustness and accuracy, the integration into real-time systems requires a careful balance between computational power, speed, and performance. The proposed method incorporates two attention modules, Body Attention and Skin Attention, that utilize contextual information to improve segmentation results. These modules draw attention to the desired areas, focusing on the body boundaries and skin pixels, respectively. Additionally, an efficient network architecture is employed in the encoder part to minimize computational power while retaining high performance. To handle the issue of noisy labels in skin datasets, the proposed method uses a weakly supervised training strategy, relying on the Skin Attention module. The results of this study demonstrate that the proposed method is comparable to, or outperforms, state-of-the-art methods on benchmark datasets.

Abstract This paper reviews the state of the art in visual privacy protection techniques, with particular attention paid to techniques applicable to the field of Active and Assisted Living (AAL). A novel taxonomy with which state-of-the-art visual privacy protection methods can be classified is introduced. Perceptual obfuscation methods, a category in this taxonomy, is highlighted. These are a category of visual privacy preservation techniques, particularly relevant when considering scenarios that come under video-based AAL monitoring. Obfuscation against machine learning models is also explored. A high-level classification scheme of privacy by design, as defined by experts in privacy and data protection law, is connected to the proposed taxonomy of visual privacy preservation techniques. Finally, we note open questions that exist in the field and introduce the reader to some exciting avenues for future research in the area of visual privacy.

As the population ages, Ambient-Assisted Living (AAL) environments are increasingly used to support older individuals’ safety and autonomy. In this study, we propose a low-cost, privacy-preserving sensor system integrated with mobile robots to enhance fall detection in AAL environments. We utilized the Luxonis OAK-D Edge-AI camera mounted on a mobile robot to detect fallen individuals. The system was trained using YOLOv6 network on the E-FPDS dataset and optimized with a knowledge distillation approach onto the more compact YOLOv5 network, which was deployed on the camera. We evaluated the system’s performance using a custom dataset captured with a robot-mounted camera. We achieved a precision of 96.52%, a recall of 95.10%, and a recognition rate of 15 frames per second. The proposed system enhances the safety and autonomy of older individuals by enabling the rapid detection and response to falls.

Privacy is a very complex concept that concerns all areas of human activity. Even more so, it has acquired whole new importance and has become a hot-button topic in the era of ubiquitous computing. The emergence of Active and Assisted Living (AAL) technologies is offering the improvement of the quality of life and maintaining the independence of older adults in their preferred environment. Especially with the advancement of computer vision, video-based technological solutions are giving us promising results but bring along significant risks of privacy violation. With this qualitative study, we aim to explore older adults’ understanding of privacy as a construct and their privacy considerations for technology-based monitoring applications in eldercare, for video-based systems among them. Exploratory interview sessions were conducted with 12 older adults 66 and plus of age living in care homes and in private households. Findings indicate that the concept of privacy is indeed difficult to understand and define for older adults. However, the context of video monitoring elicits clear privacy concerns. The nudity aspect of privacy was proved to be the most important among the participants, yet, the attitudinal shift in the perception of nudity was observed in older adults living in care homes. Eight out of 12 interviewed older adults did not like the idea of having a camera-based assistive technology, however, some of their worries were alleviated after explaining to them the possible technology’s inherent privacy-preserving techniques.

Active and Assisted Living (AAL) systems have a purpose to improve the lives of older or impaired people in various aspects. However, the use of equipment for data acquisition in these systems can be considered intrusive in some cases. Although de-identification may provide the needed protection to some extent, it is not always preferred, as it could affect the quality and utility of any obtained data. It is therefore crucial to establish methodologies for protecting the privacy of those monitored and thus affected by AAL systems. The purpose of GoodBrother is to a) analyze any issues arising from the use of monitoring AAL systems, regarding the users' privacy; b) establish proper guidelines for the use of these systems; c) develop privacy-aware methodologies for data handling; d) increase the systems' robustness and reliability; e) create databases to use towards benchmarking. Each one of these objectives are handled by separate interdisciplinary working groups.

Increasing numbers of older individuals in the societies pose great challenges for countries affected by the demographic change. The rapid development in the technological sector, on the other hand, enables various applications to make everyday life easier for older and disabled people and to maintain their autonomy for longer. This study examines the acceptance and privacy perceptions of a video-based technology for lifelogging in home environments among German and Turkish users, using a multi-method empirical research approach. Results expose an overall differing acceptance of using lifelogging cameras between German and Turkish participants and suggest that the consideration of the varying culture-bound demands is necessary. The findings of this study support the understanding of requirements for a successful implementation of a video-based assistive technology in private environments to optimally address the needs of the future users, drawing attention to the important cultural influences that affect its acceptance.

Accurate background/foreground segmentation is a preliminary process essential to most visual surveillance applications. With the increasing use of freely moving cameras, strategies have been proposed to refine initial segmentation. In this paper, it is proposed to exploit the Vide-omics paradigm, and Profile Hidden Markov Models in particular, to create a new type of object descriptors relying on spatiotemporal information. Performance of the proposed methodology has been evaluated using a standard dataset of videos captured by moving cameras. Results show that usage of the proposed object descriptors allows better foreground extraction than standard approaches.

The recognition of Activities of Daily Living (ADL) and their environments based on sensors available in off-the-shelf mobile devices is an emerging topic. These devices are capable to acquire and process the sensors' data for the correct recognition of the ADL and their environments, providing a fast and reliable feedback to the user. However, the methods implemented in a mobile application for this purpose should be adapted to the low resources of these devices. This paper focuses on the demonstration of a mobile application that implements a framework, that forks their implementation in several modules, including data acquisition, data processing, data fusion and classification methods based on the sensors? data acquired from the accelerometer, gyroscope, magnetometer, microphone and Global Positioning System (GPS) receiver. The framework presented is a function of the number of sensors available in the mobile devices and implements the classification with Deep Neural Networks (DNN) that reports an accuracy between 58.02% and 89.15%.

The need for technology assisted (or ambient assisted) living is increasing all the time as the population ages and the number of people with dementia and other conditions impairing memory and cognitive ability increases. In such applications, amongst others, it is necessary to identify and assess potentially hazardous situations. These include scenarios involving a person's hands and their interactions with various objects. In this paper, we describe our novel approach to identify human hands and objects in videos of people performing a variety of everyday tasks. We compare the performance of our method using different strategies with that of other state of the art approaches. We conclude that, when the proposed approach takes advantage of a pre-trained model, hand detection is performed accurately (94%), providing reliable information for assisted living applications.

With the development of applications associated to ego-vision systems, smart-phones, and autonomous cars, automated analysis of videos generated by freely moving cameras has become a major challenge for the computer vision community. Current techniques are still not suitable to deal with real-life situations due to, in particular, wide scene variability and the large range of camera motions. Whereas most approaches attempt to control those parameters, this paper introduces a novel video analysis paradigm, ‘vide-omics’, inspired by the principles of genomics where variability is the expected norm. Validation of this new concept is performed by designing an implementation addressing foreground extraction from videos captured by freely moving cameras. Evaluation on a set of standard videos demonstrates both robust performance that is largely independent from camera motion and scene, and state-of-the-art results in the most challenging video. Those experiments underline not only the validity of the ‘vide-omics’ paradigm, but also its potential.

The papers in this special section focus on the technologies and services that support ambient assisted living (ASL). It is now generally accepted that technology has a role to play in caring for this aging population. The concept of ambient-assisted living (AAL) is a term coined to describe the use of information and communication technologies in people’s daily living and working environment to enable them to stay active longer, remain socially connected, and live independently into old age.

Human action recognition constitutes a core component of advanced human behavior analysis. The detection and recognition of basic human motion enables to analyze and understand human activities, and to react proactively providing different kinds of services from human-computer interaction to health care assistance. In this paper, a feature-level optimization for human action recognition is proposed. The resulting recognition rate and computational cost are significantly improved by means of a low-dimensional radial summary feature and evolutionary feature subset selection. The introduced feature is computed using only the contour points of human silhouettes. These are spatially aligned based on a radial scheme. This definition shows to be proficient for feature subset selection, since different parts of the human body can be selected by choosing the appropriate feature elements. The best selection is sought using a genetic algorithm. Experimentation has been performed on the publicly available MuHAVi dataset. Promising results are shown, since state-of-the-art recognition rates are considerably outperformed with a highly reduced computational cost.

Given a territory composed of basic geographical units, the delineation of local labour market areas (LLMAs) can be seen as a problem in which those units are grouped subject to multiple constraints. In previous research, standard genetic algorithms were not able to find valid solutions, and a specific evolutionary algorithm was developed. The inclusion of multiple ad hoc operators allowed the algorithm to find better solutions than those of a widely-used greedy method. The experimentation process showed that the rate of success of each operator in generating good individuals is different and evolves with time. We therefore propose different adaptive alternatives that modify the probabilities of application of each operator throughout the evolutionary process, and compare the results of such adaptive approaches with previous results and a greedy method.

A priority two-queue system, where high-priority customers receive a variable service rate is analyzed. The model can be used as a first approximated analysis for token-ring local-area networks where voice/data integration is accomplished. The voice service rate is variable according to the queue size at the instant the server reaches the voice queue. It is assumed that the larger the number of voice packets found by the server, the higher the voice service rate is. In contrast with other approaches, instead of dropping old voice packets, the proposed scheme can be considered to drop some amount of redundant information, transmitting only the remaining part, depending on the voice queue size. This strategy provides continuity of the voice service (at the price of reducing redundancy) in such a way that the loss of quality is distributed among several voice packets, which results in more regularity in the service.>

This paper proposes a set of specific crossover and mutation operators for the delineation of functional regions through evolutionary computation. We consider a problem of dividing a given territory into local labor market areas based on spatial interaction data. Such areas are defined so that a high degree of inter-regional separation and intra-regional integration - in both cases in terms of commuting flows - exist. A genetic algorithm has been designed based on the maximization of a fitness function that measures aggregate intra-region interaction under constraints of inter-region separation and minimum size. Additional requirements, typical of any functional regionalization, include the absence of overlapping between delineated regions and an exhaustive coverage of the whole territory (so all basic spatial units must be allocated to one and only one-region). The complex set of restrictions results in conventional operators often generating invalid solutions, impeding or delaying the evolutionary process. This is the reason why an extensive set of operators has been designed that incorporates knowledge about the problem, allowing the evolution of the set of solutions towards the final result.

In this paper we study the capacities of characterization and synthesis of objects by using a self-organizing neural model, the Growing Neural Gas. These networks, by means of their competitive learning try to preserve the topology of an input space. This feature is being used for the representation of objects and their movement with topology preserving networks. We characterize the object to be represented by means of the obtained maps and kept information solely on the coordinates and the pixel color of the neurons. With this information it is made the synthesis of the original images, applying mathematical morphology and simple filters using the available information.