Abstract
The Gestalt laws of perceptual organization, which describe how visual elements in an image are grouped and interpreted, have traditionally been thought of as innate. Given past research showing that these laws have ecological validity, we investigate whether deep learning methods infer Gestalt laws from the statistics of natural scenes. We examine the law of closure, which asserts that human visual perception tends to “close the gap” by assembling elements that can jointly be interpreted as a complete figure or object. We demonstrate that a state-of-the-art convolutional neural network, trained to classify natural images, exhibits closure on synthetic displays of edge fragments, as assessed by similarity of internal representations. This finding provides further support for the hypothesis that the human perceptual system is even more elegant than the Gestaltists imagined: a single law—adaptation to the statistical structure of the environment—might suffice as fundamental.
Highlights
Psychology has long aimed to discover fundamental laws of behavior that place the field on the same footing as “hard” sciences like physics and chemistry (Schultz and Schultz 2015)
The Gestalt principle of closure asserts that human visual perception tends to “close the gap” by grouping elements that can jointly be interpreted as a complete figure or object
We compare model internal representations via a quantitative measure of closure: Ci = s(f, f) − s(f, f), where i is an index over matched image triples consisting of a complete triangle, aligned fragments, and disordered fragments; f (.) ∈ Rm is the neural net mapping from an input image in R150×150 to an mdimensional embedding, and s(., .) is a similarity function (Fig. 4)
Summary
Psychology has long aimed to discover fundamental laws of behavior that place the field on the same footing as “hard” sciences like physics and chemistry (Schultz and Schultz 2015). Perhaps the most visible and overarching set of such laws, developed in the early twentieth century to explain perceptual and attentional phenomena, are the Gestalt principles (Wertheimer 1923). These principles have had a tremendous impact on modern psychology (Kimchi 1992; Wagemans et al 2012a; Wagemans et al 2012b; Schultz and Schultz 2015). The Gestalt principle of closure asserts that human visual perception tends to “close the gap” by grouping elements that can jointly be interpreted as a complete figure or object.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
