Visual Perception: Combining Theories and Experiments to Paint a Better Picture

"Constant Curvature Segments as Building Blocks of 2D Shape Representation":

  In the research paper "Constant Curvature Segments as Building Blocks of 2D Shape Representation," Nicholas Baker and colleagues investigate the human perception of shapes, and how 2D shapes are hypothetically encoded as sets of connected constant curvature segments. They conducted this research through three separate experiments, each investigating a different aspect of their hypothesis.

    Experiment 1 used detection performance to determine whether paths with a constant curvature are easier to detect than paths with a random/alternating curvature when shown to participants among randomly orientated distractors. Their prediction about this experiment proved to be accurate when participants had a higher rate of detection of the targets with constant curvatures than the targets which lacked the constant curvature.  

    Experiment 2 broke down the perceptual analysis into segments, "by testing whether participants could learn to access hypothesized components in a contour segmentation task." This experiment found that the trial participants could learn to separate contours into segments with constant curvatures, but they could not learn to separate contours into segments with constant accelerations. These results suggest that perception allows for the mind to identify separate components of a 2D contour shape. 

    Experiment 3 investigates the representation of contours in visual processing as "built from oriented units known to exist in early cortical areas" and made to recognize constant curvature more easily. This experiment also confirmed that the visual system is more apt to detect changes to the angular extent of a segment than to detect changes to the curvature. Overall, the combination of the results from all three experiments shows that the human visual processing system uses the recognition of segments with constant curvatures as building blocks to perceiving contours. 

“Form and Function in Information for Visual Perception”:

    The article “Form and Function in Information for Visual Perception” by Joseph S. Lappin and Herbert H. Bell investigates the aspects and mechanisms of visual perception by combining James J. Gibson’s ecological approach with more contemporary theories. To begin, the article explores differing theories and definition related to the perception of information, in addition to the limitations of that perception.

    The article touches on many theories, such as Shannon's 1948 mathematical theory of communication, which defines one property of information as the relationship or correlation between two separate systems, with variations that each system presents. Essentially, information is defined as the connection between a 'sender' and a 'receiver', and cannot be viewed objectively or within the context of only one system.

    Gibson's 1979 ecological approach to information--the main theory investigated in this article--also underlined the importance of understanding the relationships within an environment in order to understand the perception of information. Perception is not necessarily the same as reality, so both must be taken into account. Gibson believed that the structural correlations between "spatiotemporal images" and objects or events in an environment provide meaning to information.

    The article also explores the differences between spatiotemporal and symbolic information, as spatiotemporal information encompasses movement and structural relations of visual images, while the theory of symbolic information breaks up images into discreet elements, or symbols. 

    Findings discovered by recent experiments lean toward the theory that visual information is spatiotemporal, and a multitude of variables within the observer's environment must be considered when analyzing visual perception. Overall, these findings suggest that vision  is able to detect spatial relations of movement more easily than stationary relations. A 2D shape, for example, is perceived through spatiotemporal structures and the relationships between them in the image of the shape. This article concludes that visual perception is most likely spatiotemporal, and that parts of the observer's environment must be taken into account when analyzing how information and objects are perceived. 

Making the Connection:

    The two research articles can be connected in multiple ways through the methods used in both to analyze and define aspects of visual perception. Baker's experiment used static symbols of curvature lines to analyze the perception of shapes. Conversely, Lappin and Bell believe that visual perception is spatiotemporal (dynamic). However, Baker's work may serve to provide evidence in support of Lappin and Bell's conclusion, because the constant curvature lines used in Baker's experiments can be seen as static representations of the dynamic forms that Lappin and Bell highlight as a major mechanism of visual perception. Essentially, Lappin and Bell provided the abstract theories and analyses behind the more quantitative data produced by Baker and his colleagues. Combining the more abstract work of Lappin and Bell's investigation of visual perception theories with the experimental data gathered by Nicholas Baker and colleagues allows for a broader understanding of the complexity of the human visual system in both processing and representing shapes and forms. For real-world use, information from both papers could be combined and used to improve image recognition and visual processing in computers and in AI programs. 

Lappin JS, Bell HH. Form and Function in Information for Visual Perception. i-Perception. 2021 Dec 23;12(6):20416695211053352. DOI: 10.1177/20416695211053352. PMID: 35003612; PMCID ;PMC8728782.

Baker N, Garrigan P, Kellman PJ. Constant curvature segments as building blocks of 2D shape representation. J Exp Psychol Gen. 2021 Aug;150(8):1556-1580. DOI: 10.1037/xge0001007. Epub 2020 Dec 17. PMID: 33332142; PMCID: PMC8324180.