返回抑制對雙穩態圖像之眼動現象的效果

Abstract

雙穩態圖像(bistable figure)是指一個能使觀察者產生兩種不同知覺的圖像，且會不由自主地在這兩種知覺間來回切換。過去已有研究指出對圖形的知覺會受到視覺注意力所聚焦之處的圖形特徵影響，而顯著性模型(saliency model)則認為視覺注意力會因返回抑制(inhibition of return)而在不同的顯著區域之間來回移動。本研究結合顯著性模型的觀點，對雙穩態圖像之知覺切換現象提出了假設：雙穩態圖像中的不同區域所包含的特徵，各自代表了不同意義，因此視覺注意力所聚焦之處決定了觀察者的知覺，而返回抑制會使視覺注意力在不同區域間遊走，使得知覺不由自主地切換。因此，本研究利用眼動追蹤儀，測量受試者觀看鴨子兔子以及奈克方塊時的視線聚焦處，並紀錄其知覺切換。在實驗一中，我們發現受試者在不同知覺下，凝視點會聚焦於畫面中的不同區域，且凝視點在不同區域間的轉移發生在知覺切換之前，表示視線聚焦之處確實決定了雙穩態圖像的知覺。在實驗二中，我們發現向內偏誤(inward bias)會影響鴨子兔子的第一知覺，但隨著觀察時間越長，這樣的偏誤會漸漸消失。而實驗三本研究則進一步遮蔽了被凝視的圖形區域，使得此區域的顯著性下降，結果發現比起實驗一，不論是知覺切換的次數還是凝視點在不同區域間的轉移次數，都有顯著的增加，表示視覺注意力的確會因畫面的顯著性改變而移動，並且造成知覺的切換。

Bistable figures can generate two different percepts and make observers’ perception spontaneously reverse. Some evidence has pointed that visual attention plays an important role in object perception because it can help us selectively focus on some features within the figure. According to the saliency model proposed by Itti and Koch (2000), visual attention can be attracted to locations of high saliency. After attention stays at the same location for a while, the local saliency will be suppressed. It thus makes visual attention shift to a new location, which is called inhibition of return. Based on this, we assumed that there are several features contained in a bistable figure, and those features imply different interpretations of the figure. Inhibition of return can make our attention shift between different regions containing those different features, and thus result in percept reversal. We used an eye-tracker to record the observers’ eye movements during observation of the duck/rabbit figure and Necker cube, and also recorded their percept reversals. In Exp. 1, we found that different fixation patterns were shown under different percepts. Also, the fixation shift across different regions occurred before the percept reversal. This supports the idea that what we perceive depends on where we look. In Exp. 2, we examined the influence of inward bias on the duck/rabbit figure. We found that it had a significant effect on the first percept, but this effect diminished over time. In Exp. 3, we added a mask to the attended region to remove the local saliency. This manipulation increased both the number of percept reversal and the number of fixation shift across different regions. That is, the change in local saliency can cause fixation shift and thus reverse our perception.