Wenfeng Chen

Image retargeting quality assessment

Content‐aware image retargeting is a technique that can flexibly display images with different aspect ratios and simultaneously preserve salient regions in images. Recently many image retargeting techniques have been proposed. To compare image quality by different retargeting methods fast and reliably, an objective metric simulating the human vision system (HVS) is presented in this paper. Different from traditional objective assessment methods that work in bottom‐up manner (i.e., assembling pixel‐level features in a local‐to‐global way), in this paper we propose to use a reverse order (top‐down manner) that organizes image features from global to local viewpoints, leading to a new objective assessment metric for retargeted images. A scale‐space matching method is designed to facilitate extraction of global geometric structures from retargeted images. By traversing the scale space from coarse to fine levels, local pixel correspondence is also established. The objective assessment metric is then based on both global geometric structures and local pixel correspondence. To evaluate color images, CIE L*a*b* color space is utilized. Experimental results are obtained to measure the performance of objective assessments with the proposed metric. The results show good consistency between the proposed objective metric and subjective assessment by human observers.

A computational cognition model of perception,memory,and judgment

The mechanism of human cognition and its computability provide an important theoretical foundation to intelligent computation of visual media. This paper focuses on the intelligent processing of massive data of visual media and its corresponding processes of perception, memory, and judgment in cognition. In particular, both the human cognitive mechanism and cognitive computability of visual media are investigated in this paper at the following three levels: neurophysiology, cognitive psychology, and computational modeling. A computational cognition model of Perception, Memory, and Judgment (PMJ model for short) is proposed, which consists of three stages and three pathways by integrating the cognitive mechanism and computability aspects in a unified framework. Finally, this paper illustrates the applications of the proposed PMJ model in five visual media research areas. As demonstrated by these applications, the PMJ model sheds some light on the intelligent processing of visual media, and it would be innovative for researchers to apply human cognitive mechanism to computer science.

Source Memory, Aging and Culture

Background: The present study investigates the possibility that culture affects age differences in context memory. There is evidence that East-Asians process scenes more holistically and show better context memory than Americans. Objective: We examined evidence for differences in binding source to context in young and old Americans and native Chinese. We hypothesized that age effects on source memory could be mitigated due to these cultural differences in processing style. Methods: During incidental encoding, younger and older Chinese and Americans watched a video with statements spoken by four distinct speakers. After a brief interval, participants identified source (experiment 1) or item and source (experiment 2). Results: We observed substantial age-related deficits in source memory in both cultures but little evidence for cultural differences in source or item memory. Conclusion: Basic source memory processes operate similarly across culture and age. The source of holistic processing differences observed between cultures may occur in cognitive operations that are more highly bound to a social context.

Identity processing in multiple-face tracking

Research has shown that observers in a multiple-object tracking task are poor at recognizing the identity of successfully tracked objects (Z. W. Pylyshyn, 2004). Employing the same paradigm, we examined identity processing and its relationship with tracking performance for human faces. Experiment 1 showed that although identity recognition was poorer after the target faces were learned in a dynamic display, identification performance was still much higher than the chance level. The experiment also found that on average about two face identities can be correctly traced to their locations. Experiment 2 showed that tracking performance decreased significantly for unique upright faces relative to the unique inverted or identical upright faces, suggesting that upright faces activate some level of mandatory identity processing that interferes and competes with visual tracking for attentional resources. Experiment 3 found that only target faces receive identity processing in the tracking task. Experiment 4 showed that switching face identities during tracking impaired tracking performance. This may indicate that identity encoding is to some extent obligatory during multiple-face tracking. Furthermore, Experiment 5 suggested that attentional resources can be consciously allocated either to maximize identity encoding or tracking, resulting in a tradeoff between the two. The results reveal a bias for face identity processing, which may differ significantly from multiple-object tracking.

Transfer between pose and expression training in face recognition

Prior research has shown that recognition of unfamiliar faces is susceptible to image variations due to pose and expression changes. However, little is known about how these variations on a new face are learnt and handled. We aimed to investigate whether exposures to one type of variation facilitate recognition in the untrained variation. In Experiment 1, faces were trained in multiple or single pose but were tested with a new expression. In Experiment 2, faces were trained in multiple or single expression but were tested in a new pose. We found that higher level of exposure to pose information facilitated recognition of the trained face in a new expression. However, multiple-expression training failed to transfer to a new pose. The findings suggest that generalisation of pose training may be extended to different types of variation whereas generalisation of expression training is largely confined within the trained type of variation.

Beauty Hinders Attention Switch in Change Detection: The Role of Facial Attractiveness and Distinctiveness

Background Recent research has shown that the presence of a task-irrelevant attractive face can induce a transient diversion of attention from a perceptual task that requires covert deployment of attention to one of the two locations. However, it is not known whether this spontaneous appraisal for facial beauty also modulates attention in change detection among multiple locations, where a slower, and more controlled search process is simultaneously affected by the magnitude of a change and the facial distinctiveness. Using the flicker paradigm, this study examines how spontaneous appraisal for facial beauty affects the detection of identity change among multiple faces. Methodology/Principal Findings Participants viewed a display consisting of two alternating frames of four faces separated by a blank frame. In half of the trials, one of the faces (target face) changed to a different person. The task of the participant was to indicate whether a change of face identity had occurred. The results showed that (1) observers were less efficient at detecting identity change among multiple attractive faces relative to unattractive faces when the target and distractor faces were not highly distinctive from one another; and (2) it is difficult to detect a change if the new face is similar to the old. Conclusions/Significance The findings suggest that attractive faces may interfere with the attention-switch process in change detection. The results also show that attention in change detection was strongly modulated by physical similarity between the alternating faces. Although facial beauty is a powerful stimulus that has well-demonstrated priority, its influence on change detection is easily superseded by low-level image similarity. The visual system appears to take a different approach to facial beauty when a task requires resource-demanding feature comparisons.

Beauty is better pursued: Effects of attractiveness in multiple-face tracking

Using the multiple-object tracking paradigm, this study examines how spontaneous appraisal for facial beauty affects distributed attention to multiple faces in dynamic displays. Observers tracked attractive faces more effectively than unattractive faces in this task. Tracking performance was only affected by target attractiveness, suggesting an absence of appraisal for distractor attractiveness. Attractive male faces also produced stronger binding of face identity and location for female participants. Together, the results suggest that facial attractiveness was appraised during tracking even though this was task irrelevant. Contrary to the theory that multiple-object tracking is driven by encapsulated low-level vision, our results show that the content of target representation is not only penetrable by social cognition but also modulates the course of tracking operations.

Matching faces with emotional expressions

There is some evidence that faces with a happy expression are recognized better than faces with other expressions. However, little is known about whether this happy-face advantage also applies to perceptual face matching, and whether similar differences exist among other expressions. Using a sequential matching paradigm, we systematically compared the effects of seven basic facial expressions on identity recognition. Identity matching was quickest when a pair of faces had an identical happy/sad/neutral expression, poorer when they had a fearful/surprise/angry expression, and poorest when they had a disgust expression. Faces with a happy/sad/fear/surprise expression were matched faster than those with an anger/disgust expression when the second face in a pair had a neutral expression. These results demonstrate that effects of facial expression on identity recognition are not limited to happy-faces when a learned face is immediately tested. The results suggest different influences of expression in perceptual matching and long-term recognition memory.

Remembering faces with emotional expressions

It is known that happy faces create more robust identity recognition memory than faces with some other expressions. However, this advantage was not verified against all basic expressions. Moreover, no research has assessed whether similar differences also exist among other expressions. To tackle these questions, we compared the effects of six basic emotional expressions on recognition memory using a standard old/new recognition task. The experiment also examined whether exposure to different emotional expressions at training creates variable effects on transfer of the trained faces to a new/neutral expression. Our results suggest that happy faces produced better identity recognition relative to disgusted faces, regardless of whether they were tested in the same image or a new image displaying a neutral expression. None of the other emotional expressions created measurable advantage for recognition memory. Overall, our data lend further support for the happy face advantage for long-term recognition memory. However, our detailed analyses also show that the advantage of happy expression on identity recognition may not be equally discernible from all other emotional expressions.

Benefits and costs of uniqueness in multiple object tracking: the role of object complexity.

How do unique objects affect multiple-object tracking? Recent research has catalogued seemingly contradictory findings, varying from enhanced to impaired tracking performance. In this study, we explore the role of object complexity in this broad range of phenomena. In a series of three experiments, we demonstrate that unique objects of varying complexity can produce both costs and benefits on tracking performance. These experiments show that the key effects of uniqueness in object tracking are results of a tradeoff between tracking operation and processing of object identity information within the capacity limit of working memory.