Perceptual filling-in is a visual illusion that occurs under certain conditions when a figure with features, such as color, brightness, and texture, all of which differ from the surrounding background, are perceived to disappear. The features are replaced, or filled in, by the surrounding background. To investigate the mechanism involved, we studied perceptual filling-in during a peripheral viewing of a gray uniform figure surrounding a dynamic random-dot texture background. Time to filling-in was measured when random-dot images with a limited bandwidth of spatio-temporal frequency were presented to subjects. On the basis of the experimental results, a hypothesis-in which filling-in arises-was proposed. This filling-in occurs because of perceptual power and temporal frequency of dynamic textures. By applying the results to the hypothesis, a decreasing rate for temporal frequency was estimated.
The head related transfer function (HRTF) method of providing spatial sound images has not been capable of dealing with individual differences between human listeners yet. The HRTF varies from person to person and with the direction of the sound source. We focus on this directional variation of the individual HRTF. Principal component analysis is applied to draw the loci of the variation in an orthogonal basis space. Each locus corresponds to a single individual HRTF, and the loci are utilized for clustering the individual HRTF. Some typical ones can be extracted from clusters; we show that they are useful in compensating for the individual differences found in HRTFs.
This paper proposes a new scheme of fusing cortex transform and brightness based features obtained by local windowing operation. Energy features are obtained by applying popular cortex transform technique within a sliding window rather than the conventional way, while we define three features namely directional surface density (DSD), normalised sharpness index (NSI), and normalized frequency index (NFI) as mesures for pixel brightness variation. Fusion by simply vector tagging as well as by correlation is performed in the feature space and then classification is done using minimum distance classifier on the fused vectors. The interesting point is that brightness features, though inferior on some natural images, often produces smoother texture boundary in mosaic images, whereas energy features show the opposite behavior.This symmetrically inverse properties are combined through vector fusion for robust classification of muti-texture images obtained from Brodatz album and VisTex database. Results and comparison in terms of edge smoothness and confusion matrix based accuracy metrics show the robustness of the scheme.
The correspondence between motion appearances and accommodative responses to two kinds of multiple-motion stimuli was examined to clarify whether the accommodation mechanism cooperates with the non-Fourier motion detection mechanism or not. The stimuli were amplitude-modulated drifting sine waves and an abruptly shifted compound sine wave that consisted of a Fourier motion component and a non-Fourier motion component moving in opposite directions. Accommodative responses were measured using an infrared optometer while observers were paying attention to each component wave's motion. The following results were obtained : (1) both the motion appearance and the accommodative state were changed depending on which component wave's motion was the object of attention; (2) each accommodative state was suitable for perceiving the attention-getting component wave's motion and separating both motion components; (3) the accommodative state shifted to one direction (far or near) as the attention-getting non-Fourier wave's spatial frequency decreased; (4) the above results were independent of the viewing distance. These results suggest that the accommodation mechanism is a cooperative system that interacts with the non-Fourier motion detection mechanism as well as the Fourier motion detection mechanism.
How to make Japanese sentences readable by adjusting both character spacing and shapes has not yet been determined. A quantitative design for creating a readable composition of Japanese short sentences is described. In this design, character spacing is quantified by a measure of distance based on the theory of induction field in vision. To find the design indices for the readable composition, we made psychological experiments using two kinds of typical Japanese sentences. As a result, an optimum composition to make these sentences readable was found. The design can be used to compose sentences of a limited line length. A practical design procedure is shown with examples for it.
A cast shadow occurs on a floor when an object occludes the illumination source from above. It is known that in a two-dimensional picture this shadow has an important role in the perception of the depth between an object and the floor. Here, we report the results of an investigation of the interaction of perceived depths performed by using binocular disparity and a cast shadow. The results rejected the winner-takes-all model where binocular disparity overrides the cast shadow. The results also indicated that both can work to produce the perception of depth. To account well for the data, we adopted a modified linear combination model for the light-source directions that subjects assume. Significant knowledge was acquired for not only bringing the visual processing of depth perception into the open, but also for creating three-dimensional computer graphics.
We studied visual fatigue in subjects after they viewed a simulated stereoscopic images with their vergence angles varied using rotary prisms to investigate whether the dissociation of vergence and accommodation is responsible for the fatigue. After approximately one hour of watching, the fatigue was assessed through subjective reports of symptoms, accommodation step responses, relative vergence limits, and visual evoked potentials (VEP). Relative vergence limits decreased and the latency of VEP increased after watching when the amount of dissociation was large and the dissociation changed over time, reflecting visual fatigue. After subjects rested, their relative vergence limits recovered.
The given time to gaze at an on-board display is very short from the viewpoint of safe driving. Therefore, the visibility of the display should be better than a general one. The appropriate method for displaying characters and symbols in an on-board display is described. In the condition of a constant surrounding illuminance, the visibility of characters and symbols depends on character complexity and viewing distance. In this method, the character complexity is shown by a two dimensional frequency (linear density). The character complexity and viewing distance are associated with visual acuity, and then the appropriate character size is determined based on the spatial frequency properties. Some experiments have showed that this method is very effective in improving display visibility.
This paper discusses navigation services utilizing 3-dimensional maps for pedestrians. We conduct two kinds of experiments.The first compares the understandability of 2- and 3-dimensional maps with different display sizes corresponding to PDAs (Personal Digital Assistants) and mobile phones. Our results indicate that 3-dimensional maps are much easier to understand than conventional 2-dimensional maps and that understandability is affected by display size.The second one deals with improving understandability of 3-dimensional maps on small display screen and it is based on the “compensation method”, which we proposed in a previous report.The compensation method adjusts parameters, such as contrast and view angle, through statistical procedures.Our findings show that the understandability of 3-dimensional maps can be improved significantly, compared with a case of usual parameter values.