Effects of 3-D Complexity on the Perception of 2-D Depictions of Objects

Abstract

In a recent study, Pelli (1999 Science 285 844-846) performed a set of perceptual experiments using portrait paintings by Chuck Close. Close's work is similar to the 'Lincoln' portraits of Harmon and Julesz (1973 Science 180 1194-1197) in that they are composite images consisting of coarsely sampled, individually painted, mostly homogeneous cells. Pelli showed that perceived shape was dependent on size, refuting findings that perception of this type is scale-invariant. In an attempt to broaden this finding we designed a series of experiments to investigate the interaction of 2-D scale and 3-D structure on our perception of 3-D shape. We present a series of experiments where field of view, 3-D object complexity, 2-D image resolution, viewing orientation, and subject matter of the stimulus are manipulated. On each trial, observers indicated if the depicted objects appeared to be 2-D or 3-D. Results for face stimuli are similar to Pelli's, while more geometrically complex stimuli show a further interaction of the 3-D information with distance and image information. Complex objects need more image information to be seen as 3-D when close; however, as they are moved further away from the observer, there is a bias for seeing them as 3-D objects rather than 2-D images. Finally, image orientation, relative to the observer, shows little effect, suggesting the participation of higher-level processes in the determination of the 'solidness' of the depicted object. Thus, we show that the critical image resolution depends systematically on the geometric complexity of the object depicted.

Perception, memory and aesthetics of indeterminate art

“…Recent functional brain imaging studies in humans have shown that object and face perception elicit activation in a distributed cortical network that encompasses a wide expanse of the visual ventral stream [10][11][12]20]. Our finding that viewers perceive recognizable objects in indeterminate paintings, in which such objects are only suggestive, is consistent with reports about the perception of ambiguous figures [16,19], illusory contours [17,28], binocular rivalry [31], and visual mental imagery [13], which showed activation in object-responsive regions in extrastriate cortex.…”

Color of Light-Source and Background Luminance on Perceived Convexity and Concavity of a Blurred 3-D Hemisphere

“…They found a preference for globally convex shapes, particularly when the stimulus was se-verely blurred. As shown by a previous study, subject matter (portraits, ecological scenes, and artificial objects), distance (the stimuli subtended approximately 6 to 60º), and resolution (4 to 81 marks placed horizontally across the widest part of the object) affect perception of 2-D and 3-D stimuli (Phillips, et al, 2004). This study reported a significant interaction between size and mark frequency, which supported Phillips, et al's (1999) findings that the size of the image in the visual field and the relative size of the marks (lines, etc.)…”

“…Another important factor in human perception of depth is the blur level of the 3-D image (Phillips, Thompson, & Voshell, 2004;Gerardin, de Montalembert, & Mamassian, 2007) or the spatial frequency of the stimulus (Goffaux & Rossion, 2006;Lee, Shioiri, & Yaguchi, 2007;Awasthi, Friedman, & Williams, 2011). Gerardin, et al (2007 used low-pass filters with different cutoff frequencies to generate different amounts of blur in stimuli to test a preference for globally convex shapes.…”

The Impacts of Multiple Robots and Display Views: An Urban Search and Rescue Simulation

“…Teleoperation of a robotic vehicle involves navigating, perceiving, understanding, and manipulating a remote environment. However, teleoperation differs from "being there in person", due to the limited field of view, distortion, low resolution of camera vision, two-dimensional perception (Phillips, Thompson, & Voshell, 2004), and constrained viewpoint. While the use of multiple views and multiple vehicles may provide assistance, limits of attentional resources, differing spatial orientations, and task switching penalties (Rubinstein, Meyer, & Evans, 2001) can cause additional confusion and limit the integration of information.…”