Are edges sufficient for object recognition?

Abstract

The authors argue that the concept of "edges" as used in current research on object recognition obscures the significant difficulties involved in interpreting stimulus information. Edges have sometimes been operationalized as line drawings, which can be an invalid and misleading practice. A new method for evaluating the utility of edge information, operationalized as the outputs of a local, signal-based edge extractor, is introduced. With 1-s exposures, the accuracy of identifying objects in the edge images was found to be less than half that with color photographs. Therefore, edges are far from being sufficient for object recognition. Alternative approaches to the problem of interpreting stimulus information are discussed.

The Structure of Three-Dimensional Object Representations in Human Vision: Evidence From Whole-Part Matching.

“…For example, in some studies, it has been shown that observers are better at recognizing gray scale images of objects (which contain information about shading and surface luminosity gradients) than line drawings (Brodie, Wallace, & Sharrat, 1991;Price & Humphreys, 1989;Sanocki et al, 1998)-although line drawings, like those used in the current study, can be sufficient to support rapid stimulus identification (Biederman & Ju, 1988).…”

Diagnostic Colors Mediate Scene Recognition

“…The name is supposed to activate a representation in memory to match against the input. When the activated representation comprises color cues, if these participate in recognition, then matching should be better with colored than with achromatic versions of the same object (Biederman & Ju, 1988;Joseph & Proffitt, 1996;Sanocki, Bowyer, Heath, & Sarkar, 1998). Studies of recognition at the basic level (Rosch, Mervis, Johnson, Grey, & BoyesBraem, 1976) revealed that subjects verified pictures of common objects equally fast whether they were colored or not (Biederman & Ju, 1988;Davidoff & Ostergaard, 1988;Ostergaard & Davidoff, 1985).…”

“…When scenes are used, it is to assess the role of perceptual context in object recognition (e.g., Aginski & Tarr, 2000;Biederman et al, 1982;Boyce, Pollatsek, & Rayner, 1989;Delorme, Fabre-Thorpe, Richard, Fize, & Thorpe, 1998;Hollingworth & Henderson, 1998;Intraub, 1997;Sanocki, Bowyer, Heath, & Sarkar, 1998). For example, Delorme et al (1998) presented humans and monkeys 400 pictures of food and animals in their natural background scene contexts.…”

Color makes a difference: Two-dimensional object naming in literate and illiterate subjects

“…On the other hand, Bierderman and Ju (1988) argue that edge-based representations are crucial for object recognition and the objects should be recognized as easily when represented by edge information as when represented by other types of information (e.g., color photos, which contain surface information such as color, texture, and relative brightness). However, according to Sanocki, Bowyer, Heath, and Sarkar (1998), using a diVerent conception of edge-based information, edge information is not suYcient for object recognition. Overall, these studies suggest that the function of color in object recognition is not well understood and there is no agreement concerning its role in object naming/recognition.…”