Gaze Sequences Research Articles

Eye Movements Reveal Optimal Strategies for Analogical Reasoning.

Analogical reasoning refers to the process of drawing inferences on the basis of the relational similarity between two domains. Although this complex cognitive ability has been the focus of inquiry for many years, most models rely on measures that cannot capture individuals' thought processes moment by moment. In the present study, we used participants' eye movements to investigate reasoning strategies in real time while solving visual propositional analogy problems (A:B::C:D). We included both a semantic and a perceptual lure on every trial to determine how these types of distracting information influence reasoning strategies. Participants spent more time fixating the analogy terms and the target relative to the other response choices, and made more saccades between the A and B items than between any other items. Participants' eyes were initially drawn to perceptual lures when looking at response choices, but they nonetheless performed the task accurately. We used participants' gaze sequences to classify each trial as representing one of three classic analogy problem solving strategies and related strategy usage to analogical reasoning performance. A project-first strategy, in which participants first extrapolate the relation between the AB pair and then generalize that relation for the C item, was both the most commonly used strategy as well as the optimal strategy for solving visual analogy problems. These findings provide new insight into the role of strategic processing in analogical problem solving.

Authoring directed gaze for full-body motion capture

We present an approach for adding directed gaze movements to characters animated using full-body motion capture. Our approach provides a comprehensive authoring solution that automatically infers plausible directed gaze from the captured body motion, provides convenient controls for manual editing, and adds synthetic gaze movements onto the original motion. The foundation of the approach is an abstract representation of gaze behavior as a sequence of gaze shifts and fixations toward targets in the scene. We present methods for automatic inference of this representation by analyzing the head and torso kinematics and scene features. We introduce tools for convenient editing of the gaze sequence and target layout that allow an animator to adjust the gaze behavior without worrying about the details of pose and timing. A synthesis component translates the gaze sequence into coordinated movements of the eyes, head, and torso, and blends these with the original body motion. We evaluate the effectiveness of our inference methods, the efficiency of the authoring process, and the quality of the resulting animation.

The Uncanny Valley and the Importance of Eye Contact

Abstract The Uncanny Valley hypothesis describes the negative emotional response of human observers that is evoked by artificial figures or prostheses with a human-like appearance. Many studies have pointed out the meaning of facial features, but did not further investigate the importance of eye contact and its role in decision making about artificial faces. In this study we recorded the number and duration of fixations of participants (N = 53) and recorded gaze movements and fixations on different areas of interest, as well as the response time when a participant judged a face as non-human. In a subsequent questionnaire, we grasped subjective ratings. In our analysis we found correlations between the likeability and the duration of eye fixations on the eye area. The gaze sequences show that artificial faces were visually processed similar to the real ones and mostly remained not assessed as artificial as long as the eye regions were not considered.

Modelling Task-Dependent Eye Guidance to Objects in Pictures

We introduce a model of attentional eye guidance based on the rationale that the deployment of gaze is to be considered in the context of a general action-perception loop relying on two strictly intertwined processes: sensory processing, depending on current gaze position, identifies sources of information that are most valuable under the given task; motor processing links such information with the oculomotor act by sampling the next gaze position and thus performing the gaze shift. In such a framework, the choice of where to look next is task-dependent and oriented to classes of objects embedded within pictures of complex scenes. The dependence on task is taken into account by exploiting the value and the payoff of gazing at certain image patches or proto-objects that provide a sparse representation of the scene objects. The different levels of the action-perception loop are represented in probabilistic form and eventually give rise to a stochastic process that generates the gaze sequence. This way the model also accounts for statistical properties of gaze shifts such as individual scan path variability. Results of the simulations are compared either with experimental data derived from publicly available datasets and from our own experiments.

Neural mechanisms for predictive head movement strategies during sequential gaze shifts

Humans adopt very different head movement strategies for different gaze behaviors, for example, when playing sports versus watching sports on television. Such strategy switching appears to depend on both context and expectation of future gaze positions. Here, we explored the neural mechanisms for such behaviors by training three monkeys to make head-unrestrained gaze shifts toward eccentric radial targets. A randomized color cue provided predictive information about whether that target would be followed by either a return gaze shift to center or another, more eccentric gaze shift, but otherwise animals were allowed to develop their own eye-head coordination strategy. In the first two animals we then stimulated the frontal eye fields (FEF) in conjunction with the color cue, and in the third animal we recorded from neurons in the superior colliculus (SC). Our results show that 1) monkeys can optimize eye-head coordination strategies from trial to trial, based on learned associations between color cues and future gaze sequences, 2) these cue-dependent coordination strategies were preserved in gaze saccades evoked during electrical stimulation of the FEF, and 3) two types of SC responses (the saccade burst and a more prolonged response related to head movement) modulated with these cue-dependent strategies, although only one (the saccade burst) varied in a predictive fashion. These data show that from one moment to the next, the brain can use contextual sensory cues to set up internal "coordination states" that convert fixed cortical gaze commands into the brain stem signals required for predictive head motion.

Repeated Web Page Visits and the Scanpath Theory: A Recurrent Pattern Detection Approach

This paper investigates the eye movement sequences of users visiting web pages repeatedly. We are interested in potential habituation due to repeated exposure. The scanpath theory posits that every person learns an idiosyncratic gaze sequence on first exposure to a stimulus and re-applies it on subsequent exposures. Josephson and Holmes (2002) tested the applicability of this hypothesis to web page revisitation but results were inconclusive. With a recurrent temporal pattern detection technique, we examine additional aspects and expose scanpaths. Results do not suggest direct applicability of the scanpath theory. While repetitive scan patterns occurred and were individually distinctive, their occurrence was variable, there were often several different patterns per person, and patterns were not primarily formed on the first exposure. However, extensive patterning occurred for some participants yet not for others which deserves further study into its determinants.

Object displays for identifying multidimensional outliers within a crowded visual periphery

This article discusses the human ability to detect, locate, or identify objects and their features using peripheral vision. The potential of peripheral vision is underused with user interfaces probably due to the limits of visual acuity. Peripheral preview can guide focused attention to informative locations, if the visual objects are large enough and otherwise within the limits of discrimination. Our experiments focused on the task of identifying an outlier and implicated another limiting factor, crowding, for integration of object features. The target object and the corresponding data dimension were located from an object display representation used for integrating multidimensional data. We measured performance on a peripheral vision task in terms of reaction times and eye movements. Subjects identified the target item from 480 alternatives within 100 ms. Therefore, the identification process would not slow down the natural gaze sequence and focused attention during monitoring and data mining tasks.

Rapid Detection of Person Information in a Naturalistic Scene

A preferential-looking paradigm was used to investigate how gaze is distributed in naturalistic scenes. Two scenes were presented side by side: one contained a single person (person-present) and one did not (person-absent). Eye movements were recorded, the principal measures being the time spent looking at each region of the scenes, and the latency and location of the first fixation within each trial. We studied gaze patterns during free viewing, and also in a task requiring gender discrimination of the human figure depicted. Results indicated a strong bias towards looking to the person-present scene. This bias was present on the first fixation after image presentation, confirming previous findings of ultra-rapid processing of complex information. Faces attracted disproportionately many fixations, the preference emerging in the first fixation and becoming stronger in the following ones. These biases were exaggerated in the gender-discrimination task. A tendency to look at the object being fixated by the person in the scene was shown to be strongest at a slightly later point in the gaze sequence. We conclude that human bodies and faces are subject to special perceptual processing when presented as part of a naturalistic scene.

Visually-guided behavior of homonymous hemianopes in a naturalistic task

The gaze behavior of homonymous hemianopes differs from that of visually intact observers when performing simple laboratory tasks. To test whether such compensatory behavior is also evident during naturalistic tasks, we analyzed the gaze patterns of three long-standing hemianopes and four visually intact controls while they assembled wooden models. No significant differences in task performance, saccade dynamics or spatial distribution of gaze were observed. Hemianopes made more look-ahead fixations than controls and their gaze sequences were less predictable. Thus hemianopes displayed none of the compensatory gaze strategies seen in laboratory tasks. Instead, their gaze patterns suggest greater updating of, and greater reliance on a spatial representation.

Eye Movements in Human Performance Modeling of Space Shuttle Operations

The goal of our research is to easily develop models that predict astronaut performance in space shuttle operations, but it is difficult to make extrapolations from astronaut training data. A solution is to decompose a complex task into a set of basic operators which are sequenced to create longer chains of behavior. In this modeling project, gaze durations and sequences are predicted and compared to the performance of novice (trained pilots) and expert (astronaut) space shuttle operators. The model makes generally good zero-parameter predictions of gaze durations, but there are notable discrepancies. The gaze sequence of the model is more similar to expert performance than novice performance, but there are differences from both. It appears that with more training, experts develop different gaze sequence strategies than novices due to familiarity with fault messages and procedures. Future modeling efforts should have their gaze sequence strategies based on expert performance.

Maternal Verbal Control Techniques with Young Infants During Mutual Gaze and Visual Co-Orientation Episodes

This study examined periods of mutual gaze and visual co-orientation between mothers and their 2-4 month-old infants as contexts for the utilization of maternal verbal control techniques; additional questions involved the impact of infant sex and maternal sensitivity upon the utilization of these techniques. Mothers and infants were videotaped during two experimental sessions, each of which was designed to enhance the occurrence of either mutual gaze or visual co-orientation. Results indicated that although a similar number of verbal control techniques occurred during mutual gaze and visual co-orientation sequences, visual context did have a significant effect upon the type of verbal control methods employed; in contrast, neither infant sex nor maternal sensitivity significantly influenced the utilization of these methods. These results demonstrate the importance of considering contextual factors when examining language/communication development; in addition, the implications of these findings for infant communication development are discussed.