The Development Of Contrast Sensitivity And The Completion Mechanism Of Stereo Illusory Contour

The perception of the world of visual system is mainly through perceptual organization, including image segmentation and grouping. For the segmentation of figure and background, one of the important information is contrast. When the luminace difference between the figure and background is bigger enough, the visual system can percept it. The function of the detection of contrast of visual system can be described as contrast sensitivity. And for the grouping, we tend to organize continuous object with the same contrast as a whole. A more sophisticated function of visual system is visual completion. Visual system fill in the part which is not shown in the visual field by generate illusory contour. Contrast sensitivity and visual completion are two fundamental topics in visual science. During my PhD, I have carried out researches in both topics.One of my work is the research of the development state of the contrast sensitivity of children. Previous studies have attempted to determine the time when contrast sensitivity becomes adult-like, while the results are conflicting. Some studies showed that contrast sensitivity was adult-like around8years old. While other studies suggested that contrast sensitivity reached adult level during adolescence. However, no study has considered the effect of higher-order aberrations (HOAs) on the measurement of contrast sensitivity. Several studies have shown that HOAs vary among individuals and affect the measurement of contrast sensitivity. Larger values of HOAs have been found in children than in adults in several studies, while their effects on contrast sensitivity of children are still not fully understood. Thus, a precise comparison that eliminates the influence of HOAs is necessary in diagnosing the status of children’s visual neural system. This study aimed to explore the neural development status of the visual system of children (around8years old) using contrast sensitivity. We achieved this by eliminating the influence of higher-order aberrations (HOAs) with adaptive optics correction.We measured HOAs, modulation transfer functions (MTFs) and contrast sensitivity functions (CSFs) of six children and five adults with both corrected and uncorrected HOAs. We found that when HOAs were corrected, children and adults both showed improvements in MTF and CSF. However, the CSF of children was still lower than the adult level, indicating different contrast sensitivity between groups cannot be explained by differences in optical factors.While the worse performance of children may simply resulted from the influence of non-neural factors, like head movement, eye movement and bad attention. These non-neural factors have been shown to increase both measured threshold and within-subject variability. Actually, we found the within-subject variability of the CSF test with AO correction were significantly larger in children than in adults. To further rule out the possible influence of non-neural factors on measurement, we conducted10-day training for children. After training, the within-subject variability of the post-training CSF measurement in children was reduced to adult level, while the contrast sensitivity of children was still worse than that of adults. This indicated the difference between the groups also could not be explained by differences in non-visual factors."With these results we concluded that the neural systems underlying vision in children of around8years old are still immature in contrast sensitivity. The research will facilitate the measurements of visual functions of children and help the detection and treatment of visual function impairment (such as amblyopia).The other work is the research of the mechanism of stereo illusory contour. Kanizsa square is a common stimulus used to study visual completion. There are two types of Kanizsa square, when perceived as sitting on top of four disks, is called modal (or illusory); when perceived as seen through four holes, it is called amodal (or occluded). The studies of whether the two visual completion have same mechanism or not got and how they are instantiated in visual circuits got controversial results. Recently, classification images have been found to be a tool to analyze visual strategies. Using classification image, researchers found that observers use perceptually interpolated contours to recognize objects. But whether the strengths of interpolation in modal and amodal contours are equally good remains unknown. To explore this, one could use three-dimensional stimulus to keep modal and amdoal stimulus same in physical input. On the other hand, considering the relationship between contour completion and binocular fusion has seldom been studied, we can also test whether contour completion occurs before or after binocular fusion by using three-dimensional stimulus.In this experiment, we investigated three-dimensional Kanizsa square discrimination from testing threshold and classification images (CIs), in order to compare modal and amodal completions and test the relationship between contour completion and binocular fusion. Cluster test was used to quantify classification images.We conducted two experiments:contrast experiment (low contrast) and angle experiment (high contrast).In both experiments, there were two noise conditions:the noise was not at the same layer as Kanizsa square; the noise was at the same layer as Kanizsa square. Our assumption1:if the completion occurs before binocular fusion, when the stimulus are physically same, even the stereo perception is different, the results of various tasks should be consistent; on the contrary, if the completion occurs after binocular fusion, and the differences of stereo perception may affect the results. Our assumption2:When the Kanizsa square is not at the same layer as noise the hypothesis is:if contour completion occurs after binocular fusion, then we would find no classification image(when the noise is at different layers with illusory contour),or classification image locate at the middle of the contours of the two monocular images,or classification image coincide with the contour of dominant-eye image; If the contour completion occurs before binocular fusion, then the classification image would locate at the contours of the two monocular images.We found that the mechanisms of modal and amodal completions are impossible to be completely same because there were different clusters between inducers and different eye dependence. When the contrast of inducers was low, the completions were mostly monocularly processed, but the process after binocular combination also affected completion. While when the contrast of inducers was high, completions happened both before and after binocular combination. Our study will help to understand the principle of visual cognitive behavior and provide theoretical support for the computer vision.