Efficient visual search by category: Specifying the features that mark the difference between artifacts and animals in preattentive vision

Levin, Daniel T.; Takarae, Yukari; Miner, Andrew G.; Keil, Frank C. · 2001 · OpenAlex-citations

DOI: 10.3758/bf03194429

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the efficiency of visual search for broad natural categories, specifically distinguishing between artifacts and animals, and identifies the perceptual features that facilitate this process. While traditional visual search research focuses on single objects defined by specific features, this work addresses how humans rapidly locate targets defined only by category membership (e.g., finding any animal among artifacts). The authors were motivated by evidence suggesting that the artifact/animal distinction is a foundational cognitive category, potentially supported by specialized perceptual processing. The study aims to determine if search for these heterogeneous categories is efficient and to specify whether this efficiency relies on complex structural descriptions or simpler global features like contour shape. The researchers conducted a series of experiments using a visual search paradigm. In Experiment 1, participants searched for a target from one category (one of 32 line drawings of animals or artifacts) among distractors from the other category in displays of three, six, or nine items. No specific target was named; subjects only knew the target category. Experiment 2 replicated this design using "jumbled" stimuli, where object parts were rearranged to disrupt structural coding while preserving basic features. Experiments 3 and 4 assessed the role of rectilinearity by correlating search slopes with the degree of straightness in object contours. Experiment 5 employed regression analyses to determine which variables—such as complexity, ambiguity, contour shape, and visual typicality—best predicted search performance. The results demonstrated that visual search by category was surprisingly efficient. In Experiment 1, search slopes for artifact targets among animals were extremely shallow (5.5–6.2 msec/item), indicating parallel processing, while animal targets among artifacts showed slightly slower but still efficient slopes (10.8–16.0 msec/item). A search asymmetry favored detecting artifacts among animals. Crucially, Experiment 2 showed that jumbling the objects did not significantly impair search efficiency; slopes remained similar to Experiment 1, suggesting that complex structural descriptions are not necessary for rapid detection. Instead, Experiments 3–5 indicated that global contour shape (rectilinearity for artifacts vs. curvilinearity for animals) and visual typicality were significant predictors of search speed. The findings conclude that efficient visual search for broad categories is driven by low-level perceptual features, particularly global contour shape, rather than high-level structural or semantic codes. This challenges theories requiring specific feature conjunctions for parallel search, demonstrating that heterogeneous categories can be parsed rapidly if they share consistent global properties. The study implies that preattentive vision utilizes these foundational perceptual distinctions to support the cognitive separation of artifacts and natural kinds, providing insight into how complex, learned categories are processed in early visual stages.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• visual search