An Explanation of Depth Effects in a Novel Illusion Through Binocular Half-Occlusion
- UNCP Author/Contributor (non-UNCP co-authors, if there are any, appear on document)
- Patrick N. Abele (Creator)
- Institution
- The University of North Carolina at Pembroke (UNCP )
- Web Site: http://www.uncp.edu/academics/library
- Advisor
- Patrick Cabe
Abstract: Three studies examined a visual depth illusion (the "bookmark illusion") and showed that it is best explained by binocular texture element occlusion-disocclusion disparity. Barrand (1979) also refers to occlusion-disocclusion disparity as "gain or loss disparity" (GOLD). Other predictions or explanations of the binocular depth illusion such as interposition, classical stereopsis, and progressive occlusion-disocclusion were generally ruled out. The illusion reverse apparent depth depending on whether viewed binocularly or monocularly. Monocularly, the shape appears to be floating above or even with the bubble sheet. Binocularly, the shape appears to be sunk into the bubble sheet. Experiment 1 quantified the magnitude of the illusion under a variety of viewing conditions with 18 observers. Observers reported whether the opaque circle appeared to be floating above, even with or sunk into the bubble sheet, in a 2 (monocular vs. binocular viewing conditions) X 3 (static, translating, or rotating movement conditions) X 3 (distances of 35, 50 or 65 cm) factorial design. Analysis of variance indicated significant main effects of viewing conditions, F(1,15)=34.752,p<.001, and for movement conditions, F(2,30)=58.308,p<.001. There was also a significant interaction between monocular/binocular viewing and movement conditions (F[2,30]=25.213,p<.001). Viewing distances had no effect on the strength of the illusion. Observers consistently reported that the circle appeared to be sunk into the bubble sheet when viewed binocularly, and even with or floating above the bubble sheet when viewed monocularly. GOLD is the only explanation that can account for the circle appearing to be sunk into the bubble sheet. Experiment 2 tested the binocular depth effect by reversing the display where the plastic "bubble sheet" was covered with black gummed paper. A hole showed a circular region of the bubble sheet. GOLD predicted that the bubble sheet should appear to bulge out over the edges of the black opaque surface around it. The 22 observers were assigned to two groups (free response or timed [t=30 sec]), and viewed the display under the same condition as Experiment 1, but only one distance of 50 cm. The procedure in a 2 (monocular vs. binocular viewing conditions) X 3 (static, translating, or rotating movement conditions) X 2 (groups; free response vs. 30 sec. timed) factorial design. Analysis of variance indicated significant main effects of viewing conditions (F[1,18]=23.867, p<.001). There was also an interaction between monocular/binocular viewing and movement conditions (F[1,18]=4.711,p<.044). The results of the experiment indicated that observers consistently reported that the bubble sheet appeared to bulge out of the surface only for the binocular viewing condition, consistent with the GOLD hypothesis. Experiment 3 tested the effects of contour orientation on depth impressions for binocular viewing. Since all explanations for binocular depth require horizontal disparity of retinal images, contours that are most vertically oriented should produce the most pronounced depth effects and contours oriented closer to horizontal should yield less pronounced depth effects. Observers should report seeing the stripe across the bubble sheet as being more sunk in as it approaches vertical and a lesser effect at stripe angles closer to horizontal. Thiry-four observers binocularly viewed a black stripe attached to the same translucent "bubble sheet" used in Experiments 1 and 2. The stripe was displaced from horizontal by various angles (0°, 22.5°, 45°, 66.5°, 90°, 112.5°, 135°, 157.5°) in seven randomized trial blocks. The design resulted in a 2 (groups; free response vs. 15 sec. timed) X2 (procedures; standard vs. modified [angle order and visual illustration of rating scale]) X 8 (angles) X 7 (trial blocks) factorial design. Analysis of variance indicated significant main effects for angles (F[7,210]=3.525,p<.001), and viewing time group (F[1,30]=24.848,p<.001). There was also an interaction between angle and trial block (F[42,1260]=9.003,p<.001). The results indicated that observers reported the stripe as being sunk deeper into the "bubble sheet" when viewed at angles closer to vertical. This is consistent with the prediction that horizontal disparity of the display would produce the most pronounced depth effect.On the whole, these studies indicated that the bookmark illusion (an opaque occluding patch seen as sunk into a special lenticular "bubble sheet" when viewed binocularly but even with or floating about the sheet when viewed monocularly) is best explained by the GOLD mechanism. The effect may be used in a number of practical applications and provide a simple means to test other aspects of the GOLD hypothesis.
An Explanation of Depth Effects in a Novel Illusion Through Binocular Half-Occlusion
PDF (Portable Document Format)
2967 KB
Created on 5/6/1996
Views: 115
Additional Information
- Publication
- Thesis
- Language: English
- Date: 1996
- Keywords
- Bookmark Illusion, Binocular Texture Element, Occlusion-disocclusion Disparity, Interposition, Stereopsis, Progressive Occlusion-disocclusion, Monocular, Binocular,