Authors: Lopamudra Choudhury, Subhrangsu Aditya, Bidyut Baran Chaudhuri
Addresses: Department of Philosophy, Jadavpur University, Kolkata 700032, India. ' Centre for Cognitive Science, Jadavpur University, Kolkata 700032 India. ' Computer Vision and Pattern Recognition Unit, Indian Statistical Institute, 203 B.T. Road, Kolkata 700108, West Bengal, India
Abstract: Muller-Lyer (ML) illusion and horizontal-vertical (HV) illusion are two of the most well-known geometrical illusions in simple line drawing. We conducted psychophysical experiments on adult human subjects using geometric shapes consisting of a combination of the components of ML-illusion as well as HV-illusion. Our findings suggest that the arrow (→) or fork (−<) junctions of ML-illusion and inverted |T| junction (⊥) of HV-illusion continue to have their implicit effects on the perceptual judgement of relative length of two line segments in their vicinity even when the task does not explicitly require our subjects to pay attention to such configurations. Cognitive psychologists have proposed perspective or depth cue based theories to explain these illusions. However, our findings seem to reveal inadequacy of these theories. Hence, we have proposed an alternative model for scene analysis that can be tested and implemented using artificial neural network and computer vision systems.
Keywords: geometrical illusion; Muller-Lyer illusion; ML illusion; horizontal-vertical illusion; HV illusion; size constancy; receptive fields; line drawing; perceptual judgement; psychophysical experiments; geometric shapes; scene analysis; artificial neural networks; ANNs; computational vision.
International Journal of Computational Vision and Robotics, 2009 Vol.1 No.1, pp.59 - 70
Published online: 26 Jun 2009 *Full-text access for editors Access for subscribers Purchase this article Comment on this article