Search

Models for image semantics understanding, such as deep learning (DL) models and mathematical models, are often trained on specific dataset or configured with specific parameters. When deploying such models on new tasks in a different test environment, it requires considerable effort to re-train the model or extensive expertise to tune the parameters. In this paper, we propose a smart reinforcement learning (RL) agent that could learn to tune parameters automatically to enhance model performance. The learning process is formulated as a generic control task for parameter adjustment, and applied to two use scenarios: (1) image attributes tuning to improve object detection performance on fixed DL model, and (2) parameter tuning of the mathematical model (Level Set) for image segmentation. We design a novel dynamic threshold mechanism in a multi-branch RL agent to effectively tune parameters of image qualities (for object detection) and Level Set models (for object segmentation). We conduct experiments on Pascal-VOC testing set, MS COCO validation set and a proprietary dataset of industrial components, where we achieve substantial improvement on object detection accuracy. We also perform experiments on the automatic parameter tuning of Level Set models. Results show that our method facilitates considerable performance improvement on public datasets compared with baseline method. [ABSTRACT FROM AUTHOR]

Production configuration is well recognised as an effective means of planning production processes for product families. The major challenge of production configuration originates from the handling of the numerous constraints associated with product and process variety. This paper develops a constraint satisfaction approach to facilitate production configuration decisions regarding constraint identification, representation, and evaluation. A domain-based model is formulated to conceptualise the production configuration process, involving inter-connections among multiple domains in conjunction with diverse domain decision variables and constraints. Within the domain framework, production configuration is formulated as a constraint satisfaction problem (CSP), which is solved using constraint heuristic search. Within constraint heuristic search, a decision propagation structure incorporating a connectionist approach is developed to facilitate the exploration of solution spaces. A case study of textile spindle production configuration is elaborated to illustrate the feasibility and potential of the domain-based CSP model for production configuration. [ABSTRACT FROM PUBLISHER]

Planning production processes for product families have been well recognised as an effective means of achieving successful product family development. However, most existing approaches do not lend themselves to planning production processes with focus on the optimality of the cohort of a product family. This paper addresses process family planning for product families. In view of the advantages of Petri nets (PNs) for modelling large systems, the potential of knowledge-based systems (KBSs) for solving complex problems and the analogy in between, we develop a methodology by integrating PNs and KBSs to support process family planning. An integrated product–process family structure, called IP2S, is proposed to organise all data pertaining to a product family and the corresponding process family, thereby anchoring planning to one platform. With the IP2S, a formal PN model of process family planning is further developed by integrating the principles of several well-defined PN extensions. Thus, this paper also contributes to visualising the dynamic behaviours and reasoning behind process family planning. The methodology is applied to process family planning for a truck family. The preliminary results demonstrate the feasibility and potential of using the methodology to support process family planning. [ABSTRACT FROM AUTHOR]

In this paper, the method based on 2-tuple linguistic is proposed to evaluate the adaptability of the emergency response plan. In this model, dynamic weights are introduced to empress different stages of the emergency. The variables are constructed to measure the correlation between the indexes in the method. Also, the indexes are constructed to evaluate the adaptability of the emergency response plan according to the characteristics of the emergency. For illustration and verification purposes, the case of sudden event from the Longjiang River cadmium pollution of Guangxi province in China is presented. Experimental results obtained demonstrate that the proposed 2-tuple linguistic methodology can improve the evaluation for the adaptability of emergency response plan. [ABSTRACT FROM AUTHOR]

With the rapid growing of wearable computing devices, indoor navigation guidance will become popular in the near future like the GPS-based navigation tools for drivers today. However, how the guided indoor navigation affects human’s memory of a novel environment has not been well studied. In this paper, we investigate route memory with three types of navigation assistance, that is, 2D map, wearable navigation assistant, and human usher. Twenty participants were asked to remember the route while being guided through a novel indoor environment. Our results show that the participants have similar patterns in remembering visual scenes, even using different types of assistance. These findings support previous work on scene memorability and provide the new insight that scene memorability is not affected by the type of navigation guidance. This may indicate that spatial working memory and visual memory are dissociated. We also show that scenes with navigation information are more memorable than scenes without such information. Finally, we provide some evidence that the location of a scene is linked to its memorability. In general, our findings provide valuable information about indoor scene memorability. [ABSTRACT FROM PUBLISHER]

Inspired by progresses in cognitive science, artificial intelligence, computer vision, and mobile computing technologies, we propose and implement a wearable virtual usher for cognitive indoor navigation based on egocentric visual perception. A novel computational framework of cognitive wayfinding in an indoor environment is proposed, which contains a context model, a route model, and a process model. A hierarchical structure is proposed to represent the cognitive context knowledge of indoor scenes. Given a start position and a destination, a Bayesian network model is proposed to represent the navigation route derived from the context model. A novel dynamic Bayesian network (DBN) model is proposed to accommodate the dynamic process of navigation based on real-time first-person-view visual input, which involves multiple asynchronous temporal dependencies. To adapt to large variations in travel time through trip segments, we propose an online adaptation algorithm for the DBN model, leading to a self-adaptive DBN. A prototype system is built and tested for technical performance and user experience. The quantitative evaluation shows that our method achieves over 13% improvement in accuracy as compared to baseline approaches based on hidden Markov model. In the user study, our system guides the participants to their destinations, emulating a human usher in multiple aspects. [ABSTRACT FROM PUBLISHER]

We present SocioGlass - a system built on Google Glass paired with a mobile phone that provides a user with in-situ information about an acquaintance in face-to-face communication. The system can recognize faces from the live feed of visual input. Accordingly, it retrieves relevant information about a person with a matching face in the database. In order to provide interaction assistance, multiple aspects of personal information are categorized based on its relevance to the interaction scenario or context. Thus, the system can be adapted to the social context in interaction assistance. The system can be used to help acquaintances build relationships, or to assist people with memory problems. [ABSTRACT FROM AUTHOR]

Dementia results in much stress in senior citizens and and immensely affects their quality of life. It also incurs huge financial and emotional burdens to their family members. Personal information assistance may alleviate such a problem by enhancing the sufferers' ability to perceive the environment, understand their personal status, and make judgments and decisions. Leveraging on visual intelligence technologies, we develop a wearable system that provides personal assistance to enhance the independence and well-being of elderly people with dementia. An application called MedHelp was developed on Google Glass paired with a mobile phone to improve medication compliance. MedHelp has three functions: it sends out reminders to a user who is on regular medication; it helps a user to recognize the correct medicine container and provides dosing instructions; and it tracks the user's medicine-taking activities so as to plan the time of next medications. [ABSTRACT FROM AUTHOR]

Computational systems for human-robot interaction (HRI) could benefit from visual perceptions of social cues that are commonly employed in human-human interactions. However, existing systems focus on one or two cues for attention or intention estimation. This research investigates how social robots may exploit a wide spectrum of visual cues for multiparty interactions. It is proposed that the vision system for social cue perception should be supported by two dimensions of functionality, namely, vision functionality and cognitive functionality. A vision-based system is proposed for a robot receptionist to embrace both functionalities for multiparty interactions. The module of vision functionality consists of a suite of methods that computationally recognize potential visual cues related to social behavior understanding. The performance of the models is validated by the ground truth annotation dataset. The module of cognitive functionality consists of two computational models that (1) quantify users' attention saliency and engagement intentions, and (2) facilitate engagement-aware behaviors for the robot to adjust its direction of attention and manage the conversational floor. The performance of the robot's engagement-aware behaviors is evaluated in a multiparty dialog scenario. The results show that the robot's engagement-aware behavior based on visual perceptions significantly improve the effectiveness of communication and positively affect user experience. [ABSTRACT FROM AUTHOR]

The paper presents a study aiming to assess how users belonging to different generations perceive service robots in the home context, in order to gain insights of multiple aspects of user needs and attitudes towards robots as to inform the design of robots that are universally acceptable. Surveys with 30 subjects from 10 Singaporean families were conducted. These subjects belonged to three generations, namely young adults, working adults, and elderly people. The survey had two phases; during the first phase, interviews were conducted in users' homes to assess their general attitudes and preferences of service robots. The second phase involved an interaction process between the users and a prototype robot in a simulated smart home environment. Users' feedback was collected according to multiple aspects of robot usability and acceptability. It was found that users of all ages were wishing for service robots to help handle home chores, and they expected the robot to complete the task with good reliability and efficiency. However, users of different age groups showed different patterns of concerns over robot usage. A multi-generational perspective is helpful to understand users' needs and attitudes that reflect age differences. It also provides insights into how members of a family share attitudes towards robots. [ABSTRACT FROM AUTHOR]

During wayfinding in a novel environment, we encounter many new places. Some of those places are encoded by our spatial memory. But how does the human brain 'decides' which locations are more important than others, and how do backtracking and repetition priming enhances memorization of these scenes? In this work, we explore how backtracking improves encoding of encountered locations. We also check whether repetition priming helps with further memory enhancement. We recruited 20 adults. Each participant was guided through an unfamiliar indoor environment. The participants were instructed to remember the path, as they would need to backtrack by themselves. Two groups were defined: the first group performed a spatial memory test at the goal destination and after backtracking; the second group performed the test only after backtracking. The mean spatial memory scores of the first group improved significantly after backtracking: from 49.8 to 60.8 %. The score of the second group was 62 %. No difference was found in performance between the first group and the second group. Backtracking alone significantly improves spatial memory of visited places. Surprisingly, repetition priming does not further enhance memorization of these places. This result may suggest that spatial reasoning causes significant cognitive load that thwarts further improvement of spatial memory of locations. [ABSTRACT FROM AUTHOR]

Scenarios have been widely used for evaluating human-robot interactions (HRIs). However, little has been reported on systematic utilization of different types of media for deploying HRI scenarios. This study investigates the methodological issues in scenario-based HRI evaluation, focusing on the effect of scenario media on user attitudes toward robots. Two experiments are designed to examine how scenario media may influence the elder adults' attitudes towards social robots. Different types of scenario media, including text, video, interactive video, and live interaction, were compared systematically with respect to established evaluation criteria. The results showed that the characteristics of scenario media influenced users' acceptance of robots and affected their attitudes. The outcome of the study helps designers to select scenario media for deploying contextual information of HRI. [ABSTRACT FROM AUTHOR]

Background: Sports and recreational facilities provide an important community resource where physical activity can be promoted among local residents. However, in China, many sports facilities are not fully open to the public and are often underutilized as essential public services. The purpose of this study was to examine the barriers to public entry from the facility management point of view.Methods: A qualitative study informed by constructivist grounded theory. Individual interviews were conducted with executive managers, marketing staff, and facilities management personnel (n = 15). We took an inductive approach to data synthesis utilizing procedural three-level coding, and used the electronic data management program (NVivo12) to organize the data analysis process.Results: The 15 participants discussed factors that impacted the sports facilities' capacity to serve the public. Four key themes emerged from data synthesis: (a) Policy-related restrictions, (b) Management-related factors, (c) Service-related factors, and (d) Supervision factors. Specific barriers affecting facilities' availability and accessibility involved policies and regulations, administrative pressure, institutional mechanisms, ideas, a professional team, practical ways, content development, schedule setting, spirit building, assessment and supervision, crisis management, and public evaluation.Conclusion: Findings from this qualitative study provided theoretical ground and empirical support for future research aimed at making sports facilities more accessible to the public in order to maximize physical activity and meet the objectives of the national fitness blueprint in China. [ABSTRACT FROM AUTHOR]

Technology development in wearable sensors and biosignal processing has made it possible to detect human stress from the physiological features. However, the intersubject difference in stress responses presents a major challenge for reliable and accurate stress estimation. This research proposes a novel cluster-based analysis method to measure perceived stress using physiological signals, which accounts for the intersubject differences. The physiological data are collected when human subjects undergo a series of task-rest cycles, incurring varying levels of stress that is indicated by an index of the State Trait Anxiety Inventory. Next, a quantitative measurement of stress is developed by analyzing the physiological features in two steps: 1) a k-means clustering process to divide subjects into different categories (clusters), and 2) cluster-wise stress evaluation using the general regression neural network. Experimental results show a significant improvement in evaluation accuracy as compared to traditional methods without clustering. The proposed method is useful in developing intelligent, personalized products for human stress management. [ABSTRACT FROM PUBLISHER]

Purpose – The purpose of this paper is twofold. First, it is to introduce a knowledge-based system for planning processes for families of final products, instead of component items, be they parts or assemblies. Second, it is to demonstrate the feasibility and potential of a prototypical system developed for planning processes families for truck families from a multinational company. Design/methodology/approach – The authors first identify the challenges in planning process families, including data and knowledge representation and constraint handling. To accommodate these challenges, the paper adopts the integrated product and process structure (IP2S) and colored timed Petri nets (CTPNs) in the proposed knowledge-based process family planning system. On top of the IP2S and CTPNs, XML-based knowledge representation is employed to alleviate the difficulties in modelling complex product and process family data and planning knowledge while enabling information exchange across different operating platforms. In addition, in accordance with the correspondence between PNs and knowledge-based systems, a mechanism is designed to cope with the generation of production rules, which model constraints. Findings – The proposed system is able to automatically generate production processes for customized products. At a higher level, such production processes provide input (e.g. operations, machines) to downstream activities for planning process details to manufacture component parts or component assemblies. Research limitations/implications – Traditional trial and error approaches to planning processes limit production performance improvement when companies need to timely produce diverse customized products. Knowledge-based systems should be developed to help companies better plan production processes based on the available manufacturing resources. Originality/value – Unlike most reported studies addressing either detailed process planning or assembly planning for component parts or component assemblies, this study tackles process planning for final products, in attempting to maintain production efficiency from a holistic view. [ABSTRACT FROM AUTHOR]

This article describes a methodology for elicitation and analysis of affective needs for vehicle design. Driven by the concept of citarasa or emotional intent, the method has five steps. First, a model of emotional intent was conceptualised; second, a semantic framework of citarasa words was developed that mapped words to specific vehicle components to form a citarasa ontology; third, customer citarasa were elicited in the field using probe interview technique; fourth, affective needs were refined through Web survey; and fifth, the elicited citarasa were analysed using data mining techniques and the citarasa analysis tool. The tool is linked to the citarasa database that enables analysis of affective needs in several countries in Europe and Asia. The system has been technically verified, validated and tested for usability with consumers and automotive end-users. [ABSTRACT FROM PUBLISHER]

Past research showed that people are able to perceive the personality of others at zero acquaintances. There are two main ways, verbal and non-verbal methods, which play an important role for one in perceiving personality of others. Extensive research was conducted in relating personality with verbal, paralinguistic and gestures cues. However, there are not much research, to our knowledge, that relates the appearance and perceived personality of robots. The main objective of this research is to relate individual design features with big five perceived personality of the robots. We used the results of rated perceptions across 100 pictorial images of robots and relate the results with the 40 individual design features using General Linear Model (GLM). The initial results of the GLM analysis showed that participants' rating of personality of robot fell along the dimension of perceived friendliness which is a common rotation of extroversion and agreeableness. Some relationships were found between humanlike design features and perceived friendliness of robots. Since participants are more familiar with humans, participants perceived robots with humanlike features friendlier than the others. Some other findings such as color and surface material were found related with participants' perceived friendliness as well. In the future, we will work on the analysis of the main and interaction effects of individual features on user's perceived friendliness. [ABSTRACT FROM AUTHOR]

Product ecosystem design entails complex user experience (UX) that involves interactions among multiple users, products, and the ambience. This paper aims to capture causal relationships between UX and design elements and in turn to provide decision support to product ecosystem analysis. A fuzzy reasoning Petri net is developed to deal with the uncertainty, complexity, and dynamics associated with UX modeling. Reasoning of diverse constructs of UX is embedded in the fuzzy production rules that are derived from self-report UX data based on rough set mining. A fuzzy reasoning algorithm is implemented to perform parallel inference by multicriteria rules and to simulate most likely UX under different ambient factors. A case study of subway station UX design demonstrates the potential of product ecosystem FRPN formulation. [ABSTRACT FROM PUBLISHER]

Recognizing the importance of user experience (UX) in product ecosystems, this paper examines the fundamental issues underlying product ecosystem design, including implications of a product ecosystem, the notion of the ambience, as well as UX in terms of affect and cognition. A conceptual model is outlined to elucidate the critical factors and the operational mechanism of product ecosystem design for user satisfaction. A technical framework of product ecosystem design for UX is presented, consisting of three consecutive and iterative stages, namely, affective-cognitive need acquisition, affective-cognitive analysis, and affective-cognitive fulfillment. An application to subway station design is reported to illustrate the key techniques for product ecosystem design, including ambient intelligence, data mining, Petri-net modeling, and simulation. [ABSTRACT FROM AUTHOR]