Perception

Unlike most multitone images, two-tone images such as print, geometric figures, and line drawings are as easy to interpret in photographic negative as in positive form. However, images derived from a multitone original in which intensity values are quantised to two levels are not. Bi-level quantised images, distinct from most other two-tone images, are shown to contain picture related components in their low spatial frequencies. Since it is the low-spatial-frequency components alone of negative images that present difficulties for vision, it is proposed that images which are as easy to interpret in negative as in positive form are those which are readily identified using only their high spatial frequencies.

Five experiments are reported in which the perception of the order of the components of line drawn patterns presented in rapid temporal sequence on a visual display unit was investigated. In experiments 1 and 2, respectively, a schematic face and an asymmetrical geometric design, and a realistic face and a symmetrical geometric design were each divided into four fragments consisting of outline and three internal features. These fragments were presented to observers in sequences in which the position of the outline in the sequence was systematically varied. Observers reported the perceived order of the fragments. If the order was misperceived the interfragment interval was increased until the sequence was correctly perceived. Analysis of the pattern of perceptual errors and the interfragment presentation interval at which the sequence was correctly perceived indicated that observers tended to perceive the sequence correctly when the outline was presented in first or last position, but had difficulty in doing this when it occupied an intermediate position. This effect was significantly stronger with facial than with geometric patterns. Furthermore, in the case of two face patterns, errors were of a form where observers reported the outline presented in positions two or three as occupying positions one or four respectively. In experiment 3 an identical procedure was used to compare the perception of temporally fragmented normal and inverted faces. The outline position effect was equally strong in both cases. In experiment 4 the relative strength of the tendency to move the outline towards first or final position was assessed by dividing each of two patterns, a face and a house, into three fragments consisting of outline and two groups of internal features. Order perception was significantly better with outline in first or third position, but where it was presented in the intermediate position it was reported as being presented in first position. In experiment 5 the general pattern of results obtained in experiments 1, 2, and 3 was verified with the use of a methodology in which pattern fragment sequences and interfragment intervals were both randomised from trial to trial and the observer's task was to specify the position of the outline in the sequence. Four patterns—a normal face, a face with inverted internal features (INF face), a face outline with irrelevant internal features (IRF face), and a geometric design—were each divided into four fragments consisting of outline and three internal features. Results indicated that for the normal and the INF face, detection of the outline position was significantly better in first and last position, whereas for the other two patterns detection was significantly better in first than in any of the other three positions. It is argued that the results of the five experiments emphasise the special importance of the outline in pattern recognition and support the existence of two processes in visual perception; a high-level parsing process and an outline priority effect. Both processes are probably precognitive in nature.

The effect of feature displacement within two well-known faces (Terry Wogan and Cyril Smith) was examined. Image processing equipment was used to produce stimuli in which the features of an original facial image were displaced to form a number of modified images. This technique was first reported by Haig, in a recognition study in which the effect of feature displacement within unfamiliar faces was investigated. In the present experiment a perceptual judgement task was carried out in which subjects were presented with a number of modified faces and asked to judge how dissimilar these were with respect to an original image. A multidimensional scaling analysis of the comparative judgements of the subjects revealed a two-dimensional solution involving displacement of the eyes and mouth. A clear division between up/down and inward/outward displacement within these features (particularly the eyes) was observed. A similar pattern of results was found for both well-known faces. This result indicates that subjects were responding to changes in the facial configuration produced by the different types of feature displacement (horizontal or vertical), as opposed to movement of the features per se. Finally, the results also indicate that the displacement of inner features (mouth, eyes, nose) was more noticeable than displacement of the outer facial features (eg hairline).

Two experiments are reported on the effect of feature masking on judgements of the sex and familiarity of faces. In experiment 1 the effect of masking the eyes, nose, or mouth of famous and nonfamous, male and female faces on response times in two tasks was investigated. In the first, recognition, task only masking of the eyes had a significant effect on response times. In the second, sex-judgement, task masking of the nose gave rise to a significant and large increase in response times. In experiment 2 it was found that when facial features were presented in isolation in a sex-judgement task, responses to noses were at chance level, unlike those for eyes or mouths. It appears that visual information available from the nose in isolation from the rest of the face is not sufficient for sex judgement, yet masking of the nose may disrupt the extraction of information about the overall topography of the face, information that may be more useful for sex judgement than for identification of a face.

Stereoacuity and its dependence on contrast were measured at four spatial frequencies separated by 1 octave steps. Using a method of adjustment, observers adjusted the retinal disparity of an aperiodic narrow-band stimulus until it appeared in the depth plane defined by two flanking reference lines. Variations in contrast affected stereoacuity (the standard deviation of ten depth settings), with better performance observed at higher contrasts. Data were fit with straight lines (on a log–log plot), indicating a power-law dependence on contrast; the slope was steeper at lower spatial frequencies. These findings are consistent with the idea that disparity is computed from the responses of size-tuned mechanisms characterized by nonlinear contrast transfer functions. In a second experiment, the effects of interocular differences in contrast on stereoacuity were studied for two conditions. In the first condition, one eye always viewed a high-contrast target while the other eye viewed targets of successively lower contrast; in the second condition, one eye always saw a target of near-threshold contrast while the other eye saw targets of successively higher contrast. When the fixed contrast was high, stereoacuity deteriorated steadily as the interocular difference in contrast increased; the loss of stereoacuity was greatest at the lowest spatial frequency. When the fixed contrast was low, however, small increases in the contrast to one eye had no deleterious effect on stereoacuity. Once interocular contrast settings exceeded a certain difference, stereoscopic acuity began to deteriorate at lower spatial frequencies. These results address the issue of the stage of visual processing at which contrast exerts its influence on stereopsis.

A simple scheme for deriving relative depth (time-to-collision, or TTC) from optic flow is developed in which the total flow is first sensed by unconnected motion (imperfect filter) sensors and then the rotational component is subtracted to yield the translational component. Only the latter component yields depth information. This scheme is contrasted with one where the TTC sensors respond only to the translational component at the initial registration of the flow (perfect filter sensors or looming detectors). The simple scheme predicts the results of three experiments on discrimination of TTC: discrimination thresholds are elevated if the objects withdraw from rather than approach the observer, thresholds are elevated if a rotational component is added to the flow, and the amount of threshold elevation resulting from the addition of a rotational component is reduced by prior adaptation to a pure rotational flow. These results confirm the simple model and disconfirm predictions based on the looming detector scheme.

The causal flows between the processes responsible for illusory contour clarity, brightness, and apparent depth in the Kanizsa square were examined. The sixty-four stimuli used consisted of all possible combinations of eight disk luminances and eight centre-to-centre separations between nearest disks. Ten subjects were instructed to rate the clarity of the illusory contour and the brightness and apparent depth differences between the Kanizsa square and its surround in each stimulus. On the basis of results obtained with the causal inference method, using partial correlations and path analysis, it is suggested that clarity of illusory contour can be influenced directly by disk separation, and that the output from the process responsible for illusory contour clarity has some effect on the processes responsible for the apparent depth and brightness differences.

If a line (the pointer) is aligned with a dot (the target) that stands on another line (the induction line) which is at an angle to the pointer, the pointer and the dot may no longer appear collinear. Whether they do or not depends upon the angle formed by the pointer with the induction line: the smaller the angle, the greater the misalignment effect. Misalignment is always in the direction of the induction line, which is why this alignment illusion is called attraction-caused misalignment (attraction misalignment for short). Three experiments are described in which this illusion is explored further. In the first it is shown that the induction line can exert its influence even when not contiguous with the target, though the size of the effect varies inversely with the distance of the induction line from the target. In the second experiment it is demonstrated that a dot as well as a line can induce attraction misalignment and that similarity between the induction and target items increases misalignment. Evidence in support of the theory that the termination of the induction line, as well as the part contiguous with the target dot, may induce attraction misalignment is provided in the third experiment.

In order to evaluate the importance of the axis of stimulus presentation, inter- and intramanual recognition of mirror pairs was studied with the stimulus materials aligned along the front/back axis (whereas in previous work the mirror pairs were aligned along the left/right axis). Children were allowed to feel shapes with the whole hand, with only four fingers (excluding the thumb), or with only the index finger. After learning with one hand, recognition was tested in experiment 1 with the other hand; after learning with one orientation of the hand (palm down or up), recognition was tested in experiment 2 with the other orientation (palm up or down) of the same hand; after learning with one coronal alignment of the hand (to the left or right), recognition was tested in experiment 3 with the other alignment (to the right or left), but without rotation, of the same hand. Significantly fewer intermanual recognition errors were made on mirror pairs with the materials oriented along the front/back axis than in previous work when oriented along the left/right axis. This supports the suggestion that such errors arise when the stimuli are oriented along the left/right axis during formation of the memory trace. The same trend was unexpectedly obtained for intramanual recognition errors (after rotation of the hand). These errors (after hand rotation) are largely due to coding with respect to the hand; they are reduced when the hand is not aligned with the body axis, since then coding can also occur in relation to the environment.

In reply to a suggestion made by Galton in 1875, the ability of dogs to discriminate between the odour of human twins was investigated. In a matching-to-sample simultaneous discrimination task, dogs were tested on their ability to discriminate odours from twins differing only in genetic relatedness or only in environmental factors, particularly diet, or from twins identical in both genetic relatedness and environmental factors. Dogs could discriminate between the odours from twins who differed only in environmental factors and between the odours of twins who differed only in genetic relatedness. However, they were unable to discriminate between odours produced by infant twins identical in both genetic relatedness and environmental factors. Thus twins may be discriminated by dogs as long as they differ in genetic relatedness or environmental factors. The possible source of discriminable odours is discussed and how the effects of genes and environment are mediated considered.