The validated circular shape space: Quantifying the visual similarity of shape.

Abstract

Subjective similarity holds a prominent place in many psychological theories, influencing diverse cognitive processes ranging from attention and categorization to memory and problem solving. Despite the known importance of subjective similarity, there are few resources available to experimenters interested in manipulating the visual similarity of shape, one common type of subjective similarity. Here, across seven validation iterations, we incrementally developed a stimulus space consisting of 360 shapes using a novel image-processing method in conjunction with collected similarity judgments. The result is the Validated Circular Shape space, the first Validated Circular Shape space comparable to the commonly used “color wheel”, whereby angular distance along a 2D circle is a proxy for visual similarity. This extensively validated resource is freely available to experimenters wishing to precisely manipulate the visual similarity of shape.

An image-computable model of human visual shape similarity

“…Li et al [53] constructed a 'perceptually circular' stimulus set, which ShapeComp predicts quite well (Figure S7).…”

“…Many previous studies have sought to measure shape similarity for both familiar and unfamiliar objects [38,40,[44][45][46][47][48][49][50][51][52][53]. Despite this, the representation of shape in the human visual system remains elusive, and the basis for shape similarity judgments remains unclear.…”

Visual interference can help and hinder memory: Capturing representational detail using the Validated Circular Shape Space

“…In order to study how interference impacts shape memory, we previously created the Validated Circular Shape Space (VCS space), a circular space whereby angular distance on a 2D circle is a proxy for visual similarity (Li, Liang, Lee, & Barense, 2020). VCS space was designed and validated to possess two properties ( Figure 1b): 1) it is perceptually uniform, meaning that two shapes separated by any distance are about as visually similar as any two other shapes separated by the same distance, and 2) angular distance along the circle approximates the magnitude of visual similarity between two points, meaning that shapes sampled closer in VISUAL INTERFERENCE CAN HELP AND HINDER MEMORY 7 angular distance are more visually similar than shapes sampled further apart in angular distance.…”

“…This debate was reignited with the observation that angular distance on circular space approximates the magnitude of visual similarity but is not identical to visual similarity (Schurgin, Wixted, & Brady, 2019), meaning that angular distance does not reflect a one-to-one correspondence with the magnitude of visual similarity between two points on the circle. For example, although two items sampled from 60 degrees apart on the circular space are more visually similar than two items sampled from 120 degrees apart, the two items sampled from 60 degrees apart are not necessarily twice as similar as items sampled from 120 degrees ( Figure 1b; Li et al, 2020). Importantly, this nonlinear relationship between the angular distance and the magnitude of visual similarity between two points on circular space may influence the interpretation of the mixture model, as the uniform distribution is fit to the data with the assumption that all errors on circular space are equally likely during a random guess (Figure 2a).…”

“…We directly accounted for the relationship between angular distance and the magnitude of visual similarity, comparing the results obtained from the mixture model to a parameter-free analysis that did not fit the mixture model. Using shapes whose visual similarity was explicitly quantified by VCS space (Li et al, 2020), we found that dissimilar interference disrupted both fine-and coarse-grained information, akin to memory erasure. In contrast, similar interference disrupted fine-grained but increased the reliance on coarse-grained information, akin to memory blurring.…”