Synthetic aperture radar (SAR) images are often contaminated by speckle noise, a type of multiplicative noise resulting from the imaging process. SAR image despeckling is a crucial preprocessing step for satellite imaging, enhancing image visualization and facilitating downstream analysis. In this study, we propose a linear–angular attention transformer network for SAR despeckling. The approach efficiently captures both local and global context in linear time within a multiscale transformer-convolutional neural network architecture. Our transformer integrates nonlocal denoising and multiscale feature extraction in a single model. Using a smoothing loss and fast nonlocal post-processing, the model achieved a 17% improvement in structural similarity index, a 61% enhancement in contrast-to-noise ratio, and an increase above 100% in the equivalent number of looks metric across three datasets compared to multiple state-of-the-art baselines, even when trained on a small dataset and for only 13 epochs. Comparison with different multi-head self-attention mechanisms revealed the effectiveness of linear–angular attention as a step towards green AI, showing both quantitative and qualitative performance improvements. Unlike models that rely on optical images for training and lack domain-specific features for real SAR despeckling, the proposed network is trained directly on SAR images in a self-supervised manner.

The uncanny valley (UV) effect is a negative affective reaction to human-looking artificial entities. It hinders comfortable, trust-based interactions with android robots and virtual characters. Despite extensive research, a consensus has not formed on its theoretical basis or methodologies. We conducted a meta-analysis to assess operationalizations of human likeness (independent variable) and the UV effect (dependent variable). Of 468 studies, 72 met the inclusion criteria. These studies employed 10 different stimulus creation techniques, 39 affect measures, and 14 indirect measures. Based on 247 effect sizes, a three-level meta-analysis model revealed the UV effect had a large effect size, Hedges’ g = 1.01 [0.80, 1.22]. A mixed-effects meta-regression model with creation technique as the moderator variable revealed face distortion produced the largest effect size, g = 1.46 [0.69, 2.24], followed by distinct entities, g = 1.20 [1.02, 1.38], realism render, g = 0.99 [0.62, 1.36], and morphing, g = 0.94 [0.64, 1.24]. Affective indices producing the largest effects were threatening, likable, aesthetics, familiarity , and eeriness , and indirect measures were dislike frequency, categorization reaction time, like frequency, avoidance , and viewing duration . This meta-analysis—the first on the UV effect—provides a methodological foundation and design principles for future research.

The ability of computers to recognise hand gestures visually is essential for progress in human–computer interaction. Gesture recognition has applications ranging from sign language to medical assistance to virtual reality. However, gesture recognition is extremely challenging not only because of its diverse contexts, multiple interpretations, and spatio‐temporal variations but also because of the complex non‐rigid properties of the hand. This study surveys major constraints on vision‐based gesture recognition occurring in detection and pre‐processing, representation and feature extraction, and recognition. Current challenges are explored in detail.

BACKGROUND: Over the past decade, pathway and gene-set enrichment analysis has evolved into the study of high-throughput functional genomics. Owing to poorly annotated and incomplete pathway data, researchers have begun to combine pathway and gene-set enrichment analysis as well as network module-based approaches to identify crucial relationships between different molecular mechanisms. METHODS: To meet the new challenge of molecular phenotype discovery, in this work, we have developed an integrated online database, the Pathway And Gene Enrichment Database (PAGED), to enable comprehensive searches for disease-specific pathways, gene signatures, microRNA targets, and network modules by integrating gene-set-based prior knowledge as molecular patterns from multiple levels: the genome, transcriptome, post-transcriptome, and proteome. RESULTS: The online database we developed, PAGED http://bio.informatics.iupui.edu/PAGED is by far the most comprehensive public compilation of gene sets. In its current release, PAGED contains a total of 25,242 gene sets, 61,413 genes, 20 organisms, and 1,275,560 records from five major categories. Beyond its size, the advantage of PAGED lies in the explorations of relationships between gene sets as gene-set association networks (GSANs). Using colorectal cancer expression data analysis as a case study, we demonstrate how to query this database resource to discover crucial pathways, gene signatures, and gene network modules specific to colorectal cancer functional genomics. CONCLUSIONS: This integrated online database lays a foundation for developing tools beyond third-generation pathway analysis approaches on for discovering molecular phenotypes, especially for disease-associated pathway/gene-set enrichment analysis.

Culturally influenced preferences in website aesthetics is a topic often neglected by scholars in human–computer interaction. Kim, Lee, and Choi (2003 Kim, J., Lee, J. and Choi, D. 2003. Designing emotionally evocative homepages: An empirical study of the quantitative relations between design factors and emotional dimensions. International Journal of Human–Computer Interaction Studies, 59: 899–940. [Crossref], [Web of Science ®] , [Google Scholar]) identified aesthetic design factors of web home pages that elicited particular responses in South Korean web users based on 13 secondary emotional dimensions. This study extends Kim et al.'s work to U.S. participants, comparing the original South Korean findings with U.S. findings. Results show that U.S. participants reliably applied translations of the emotional adjectives used in the South Korean study to the home pages. However, factor analysis revealed that the aesthetic perceptions of U.S. and South Korean participants formed different aesthetic dimensions composed of different sets of emotional adjectives, suggesting that U.S. and South Korean people perceive the aesthetics of home pages differently. These results indicate that website aesthetics can vary significantly between cultures.

The novice–expert ratio method (NEM) pinpoints user interface design problems by identifying the steps in a task that have a high ratio of novice to expert completion time. This study tested the construct validity of NEM's ratio measure against common alternatives. Data were collected from 337 participants who separately performed 10 word-completion tasks on a cellular phone interface. The logarithm, ratio, Cohen's d, and Hedges's ĝ measures had similar construct validity, but Hedges's ĝ provided the most accurate measure of effect size. All these measures correlated more strongly with self-reported interface usability and interface knowledge when applied to the number of actions required to complete a task than when applied to task completion time. A weighted average of both measures had the highest correlation. The relatively high correlation between self-reported interface usability and a weighted Hedges's ĝ measure as compared to the correlations found in the literature indicates the usefulness of the weighted Hedges's ĝ measure in identifying usability problems.

The eerie feeling attributed to human-looking robots and animated characters may be a key factor in our perceptual and cognitive discrimination of the human and humanlike. This study applies regression, the generalized linear model (GLM), factor analysis, multidimensional scaling (MDS), and kernel isometric mapping (Isomap) to analyze ratings of 27 emotions of 18 moving figures whose appearance varies along a human likeness continuum. The results indicate (1) Attributions of eerie and creepy better capture our visceral reaction to an uncanny robot than strange. (2) Eerie and creepy are mainly associated with fear but also shocked, disgusted, and nervous. Strange is less strongly associated with emotion. (3) Thus, strange may be more cognitive, while eerie and creepy are more perceptual/emotional. (4) Human features increase ratings of human likeness. (5) Women are slightly more sensitive to eerie and creepy than men; and older people may be more willing to attribute human likeness to a robot despite its eeriness.

Technology has enabled mega-collaboration on an unprecedented scale. A tool is needed to coordinate these activities and link them to government response efforts. However, in defining and responding to problems, teammates need to be able to visualize each other's mental models. The tool must encourage the team to advance promptly through team "forming, norming, storming, and performing." Finally, it must help individual teams visualize efficiently the "big picture," using agents to enhance the process. The usability of a prototype mega-collaboration interface has been tested, the second generation interface is being implemented, and a novel agent-based interface is being planned.

The human body constructs itself into a person by becoming attuned to the affective consequences of its actions in social relationships. Norms develop that ground perception and action, providing standards for appraising conduct. The body finds itself motivated to enact itself as a character in the drama of life, carving from its beliefs, intentions, and experiences a unique identity and perspective. If a biological body can construct itself into a person by exploiting social mechanisms, could an electromechanical body, a robot, do the same? To qualify for personhood, a robot body must be able to construct its own identity, to assume different roles, and to discriminate in forming friendships. Though all these conditions could be considered benchmarks of personhood, the most compelling benchmark, for which the above mentioned are prerequisites, is the ability to sustain long-term relationships. Long-term relationships demand that a robot continually recreate itself as it scripts its own future. This benchmark may be contrasted with those of previous research, which tend to define personhood in terms that are trivial, subjective, or based on assumptions about moral universals. Although personhood should not in principle be limited to one species, the most humanlike of robots are best equipped for reciprocal relationships with human beings

Could androids use movements to build relationships? For people, relationships are created with the help of behaviour-shaping norms, which infants begin to discover and manipulate by the third month. To build relationships, MAchines can also learn to exploit human reactions in real-time decision-making. In the video game Tetris, for example, affect co-opts computer-generated patterns to simplify cognitive tasks: norms mediate what Kirsh and Maglio (Cognitive Sci., 18, pp. 513–549, 1994) term epistemic actions, which allow implicit knowledge to shape key pressing in ways that, given past games, are likely to be informative and valuable. Expert players act to change their cognitive states by allowing the game's higher-level states to constrain their lower-level actions. Since this process enables the development of expertise, we might expect it to be widespread; but it seems marginal in hamadryas baboons, although they use affect and complex norms. In humans, by contrast, infants use adults as cognitive resources in developing their epistemic abilities. This has engineering implications for android designers. Since androids can elicit epistemic actions, engineers need to develop an affect–sensitive interface. If successful at this, even rudimentary co-action may prompt people to report experiencing androids as both making choices and violating expectations.

Android science is a new interdisciplinary framework for studying human cognition and interaction based on the premise that a very humanlike robot can elicit the sort of responses people typically ...

One of the main aims of humanoid robotics is to develop robots that are capable of interacting naturally with people. However, to understand the essence of human interaction, it is crucial to investigate the contribution of behavior and appearance. Our group's research explores these relationships by developing androids that closely resemble human beings in both aspects. If humanlike appearance causes us to evaluate an android's behavior from a human standard, we are more likely to be cognizant of deviations from human norms. Therefore, the android's motions must closely match human performance to avoid looking strange, including such autonomic responses as the shoulder movements involved in breathing. This paper proposes a method to implement motions that look human by mapping their three-dimensional appearance from a human performer to the android and then evaluating the verisimilitude of the visible motions using a motion capture system. This approach has several advantages over current research, which has focused on copying a person's moving joint angles to a robot: (1) in an android robot with many degrees of freedom and kinematics that differs from that of a human being, it is difficult to calculate which joint angles would make the robot's posture appear similar to the human performer; and (2) the motion that we perceive is at the robot's surface, not necessarily at its joints, which are often hidden from view.

Artificial neural networks have been traditionally employed to learn and compute the inverse kinematics of a robotic arm. However, the inverse kinematics model of a typical robotic arm with joint limits is a multi-valued and discontinuous function. Because it is difficult for a multilayer neural network to approximate this type of function, an accurate inverse kinematics model cannot be obtained by using a single neural network. In order to overcome the difficulties of inverse kinematics learning, we propose a novel modular neural network system that consists of a number of expert modules, where each expert approximates a continuous part of the inverse kinematics function. The proposed system selects one appropriate expert whose output minimizes the expected position/orientation error of the end-effector of the arm. The system can learn a precise inverse kinematics model of a robotic arm with equal or more degrees of freedom than that of its end-effector. However, there are robotic arms with fewer degrees of freedom, where the system cannot learn their precise inverse kinematics model. We have adopted a modified Gauss-Newton method for finding the least-squares solution to address this issue. Through the modifications presented in this paper, the improved modular neural network system can obtain a precise inverse kinematics model of a general robotic arm.

Developing a humanoid robot that can learn to perform complex tasks by itself has become a major goal of robotics research. This paper proposes a new algorithm for the automatic segmentation of humanoid motion data. We use a simplified soccer game, RoboCup, as a prototype task. Motion data from a 20 degree-of-freedom humanoid soccer playing robot are reduced to their intrinsic dimensionality by nonlinear principal component analysis. The proposed algorithm operates in of two phases. The first phase automatically segments the motion data in the reduced sensorimotor space by incrementally generating nonlinear principal component analysis with a circular constraint networks and assigning data points based on their temporal order to these networks in a conquer-and-divide fashion. Then, the second phase of the algorithm removes repeated patterns based on the distance between redundant motion patterns in the reduced sensorimotor space. The networks abstracted five motion patterns without any prior information about the number or type of motion patterns. The learned networks can be used to recognize and generate humanoid actions.

The goal of this research is to develop systems that support human activities through environmental intelligence. These systems may include agents that interact with people in a sensor-rich setting. As a first step, we have developed a distributed vision system with 12 omnidirectional cameras to recognize what is happening at a subway station. Conventional recognition systems require intermediate models of human behavior, such as geometrical models. However, it is hard to calibrate many cameras with enough accuracy to use this framework. Therefore, we propose a memory-based method that does not require any calibration or a priori models. The developed system recognizes "people are walking toward a ticket gate," "people are going down the stairs," and so on. In addition, we have developed a method to select what information is necessary to make these discriminations. This method, which is analogous to an aspect of attention, reduces the computation involved in learning support vector machines to about 0.5% of its original value and increases classification accuracy by 2 to 23%.

This paper explores memory-based approaches to the recognition of human behavior that relies on a database of previously categorized instances of sensory data. To overcome the curse of dimensionality, we examine two related methods that both rely on a hierarchical division of the sensory space using a decision tree. The first approach iteratively applies linear discriminant analysis to divide the sensory space in half in order to construct a binary tree for recognizing behaviors. We have verified the effectiveness of this approach for real-time behavior recognition using infrared sensors distributed in a desk environment and compared its results to those of Quinlan's C4.5. The second approach applies the well-known ID3 algorithm to the construction of a decision tree based on an information criterion. We use it to recognize browsing behavior at a video rental shop. Inferences are derived directly from the binarized pixel data of four wide-view cameras. Both systems offer behavior recognition rates in excess of 90%.

An essential factor in understanding the motor learning capability of humans, is the coordinate transformation learning of the visual feedback controller. Although a number of learning models for the visual feedback controller have been proposed, none has been able to establish a definitive approach. In our previous work, we have suggested a learning model that uses disturbance noise and the feedback error signal to learn the human visual feedback controller's coordinate transformation. However, the model does not fully consider the time delay in the visual feedback control loop. This paper presents a modified learning model taking into account the time delay and the convergence properties of the model. Numerical simulations are presented to illustrate the effectiveness of the proposed approach.

Robotics can serve as a testbed for cognitive theories. One behavioral criterion for comparing theories is the extent to which their implementations can learn to exploit new environmental opportunities. Furthermore, a robotics testbed forces researcher to confront fundamental issues concerning how internal representations are grounded in activity. In our approach, a mobile robot takes the role of a creature that must survive in an unknown environment. The robot has no a priori knowledge about what constitutes a suitable goal-what is edible, inedible, or dangerous-or even its shape or how its body works. Nevertheless, the robot learns how to survive. The robot does this by tracking segmented regions of its camera image while moving. The robot projects these regions into a canonical wavelet domain that highlights color and intensity changes at various scales. This reveals sensory invariance that is readily extracted with Bayesian statistics. The robot simultaneously learns an adaptable sensorimotor mapping by recording how motor signals transform the locations of regions on its camera image. The robot learn about its own physical extension when it touches an object. But it also undergoes an internal state change analogous to the thirst quenching or nausea producing effects of intake in animals. This allows the robot to learn what an object affords by relating these effects to learned clusters of invariance. In this way primitive symbols emerge. These protosymbols provide the robot with goals that it can achieve by using its sensorimotor mapping to navigate, for example, toward food and away from danger.

The speed, accuracy, and adaptability of human movement depends on the brain performing an inverse kinematics transformation-that is, a transformation from visual to joint angle coordinates-based on learning from experience. In human motion control, it is important to learn a feedback controller for the hand position error in the human inverse kinematics solver. This paper proposes a novel model that uses disturbance noise and the feedback error signal to learn coordinate transformations of the human visual feedback controller. The proposed model redresses drawbacks in current models because it does not rely on complex signal switching, which does not seem neurophysiologically plausible. Numerical simulations show the effectiveness of the model.

An important aim of robotics is to design and build machines that can recognize and exploit opportunities afforded by new situations. Traditionally in artificial intelligence, this task has fallen on abstract representations, but that has left the problem of how to ground the representations in sensorimotor activity. The paper proposes a computational architecture whereby a mobile robot internalizes representations based on its experience. The paper examines a fast online learning algorithm that allows the robot to build up a mapping of how its motor signals transform sensory data. The author also proposes a way of categorizing object affordances according to their internal effects. Based on these effects, wavelet analysis is applied to sensory data to uncover invariance, thus abstracting a representation of goals. Finally, heuristics are explored for projecting a learned sensorimotor mapping into the future to attain these goals.

Robotics can serve as a testbed for cognitive theories. One behavioral criterion for comparing theories is the extent to which their implementations can learn to exploit new environmental opportunities. Furthermore, a robotics testbed forces researchers to confront fundamental issues concerning how internal representations are grounded in activity. In our approach, a mobile robot takes the role of a creature that must survive in an unknown environment. The robot has no a priori knowledge about what constitutes a suitable goal, what is edible, inedible, or dangerous, or even its shape or how its body works. Nevertheless, the robot learns how to survive. The robot does this by tracking segmented regions of its camera image while moving. The robot projects these regions into a canonical wavelet domain that highlights color and intensity changes at various scales. This reveals sensory invariance that is readily extracted with Bayesian statistics. The robot simultaneously learns an adaptable sensorimotor mapping by recording how motor signals transform the locations of regions on its camera image. The robot learns about its own physical extension when it touches an object, but it also undergoes an internal state change analogous to the thirst quenching or nausea producing effects of intake in animals. This allows the robot to learn what an object affords: is it edible or poisonous, by relating these effects to learned clusters of invariance. In this way primitive symbols emerge. These proto-symbols provide the robot with goals that it can achieve by using its sensorimotor mapping to navigate, for example, toward food and away from danger.

In the drive towards miniaturization in manufacturing, accuracy in positioning minute objects by camera is vital. For visual servoing, the rapid and robust detections of features in camera images is also essential to production line efficiency. Template matching provides flexibility in achieving this, often lacked by other methods, because it avoids the need to set object-specific parameters. Unfortunately, standard methods of template matching require much calculation, especially for detecting feature rotation. The delay this causes means that for many applications template matching provides too slow a source of visual feedback. As an alternative, we propose a new method of detecting the translation and rotation of a feature from coarse optical flow, we and apply it to visual servoing. Coarse optical flow is derived from the difference in intensity between a region of the initial and current image and their pixel-by-pixel intensity gradients. Unlike template matching, our method can detect large rotations with relatively little calculation. Image resolution is then adjusted from coarse to fine. Subpixel accuracy results in a 100 fold improvement in precision (by area). We show experimental results for precise planar positioning.