Scene Consistency in Object and Background Perception

Abstract

Does knowledge about which objects and settings tend to co-occur affect how people interpret an image? The effects of consistency on perception were investigated using manipulated photographs containing a foreground object that was either semantically consistent or inconsistent with its setting. In four experiments, participants reported the foreground object, the setting, or both after seeing each picture for 80 ms followed by a mask. In Experiment 1, objects were identified more accurately in a consistent than an inconsistent setting. In Experiment 2, backgrounds were identified more accurately when they contained a consistent rather than an inconsistent foreground object. In Experiment 3, objects were presented without backgrounds and backgrounds without objects; comparison with the other experiments indicated that objects were identified better in isolation than when presented with a background, but there was no difference in accuracy for backgrounds whether they appeared with a foreground object or not. Finally, in Experiment 4, consistency effects remained when both objects and backgrounds were reported. Semantic consistency information is available when a scene is glimpsed briefly and affects both object and background perception. Objects and their settings are processed interactively and not in isolation.

On the functions, mechanisms, and malfunctions of intracortical contextual modulation

“…Effects of these contextual biases range from facilitation of object-naming at the subordinate level (Davenport and Potter, 2004) to the rapid detection of super-ordinate object categories such as animals (Fize, Cauchoix and Fabre-Thorpe, 2011). Furthermore, the latter study also showed that the facilitation of object recognition by learned associations between object and background scene is very similar in humans and macaques.…”

Speed limits: Orientation and semantic context interactions constrain natural scene discrimination dynamics.

“…In this experiment, we manipulated the semantic congruency of the meaningful objects and scenes by varying the likelihood of their co-occurrence, one aspect of the semantic relation between them (for a discussion of other aspects, see Biederman et al, 1982). This approach is similar to that used in previous studies using line drawings of objects and contexts (Biederman et al, 1982;Hollingworth & Henderson, 1998) and natural scene photos (Davenport & Potter, 2004). In pictures with semantically congruent object-context relation, it was very likely for the participants that the object could be found in the particular scene context (e.g., deer on grassland; see Figure 1G), whereas in pictures with incongruent semantic object-context relations, the objects were presented in an atypical surrounding scene context (e.g., a bookshelf on a beach; see Figure 1G).…”

“…A newer study by Davenport and Potter (2004) using natural scene photographs reported that semantic scene congruency has an effect on the processing of both the target object and its context. It is interesting to note that these investigators found that accuracy for objects presented on a uniform white background was better than for objects presented on either a congruent or an incongruent background.…”

The importance of information localization in scene gist recognition.

“…The term gist is not always clearly defined (though see Oliva, 2005) but is most frequently operationalized as the scene's basic level category, for example, "beach" or "street" (Tversky & Hemenway, 1983), and we follow that convention here. Gist information appears to guide viewers' inspection of the scene (Loftus & Mackworth, 1978;Oliva, Torralba, Castelhano, & Henderson, 2003), may aid object recognition in the scene (Boyce & Pollatsek, 1992;Davenport & Potter, 2004;De Graef, De Troy, & D'Ydewalle, 1992;Hollingworth & Henderson, 1998;Palmer, 1975), and affects later memory of the scene (Brewer & Treyens, 1981;Pezdek, Whetstone, Reynolds, Askari, & Dougherty, 1989). Given the speed of gist perception, the information underlying gist recognition may be based on holistic, low-level scene properties (Oliva & Torralba, 2001;Renninger & Malik, 2004;Vailaya, Jain, & Zhang, 1998), rather than based on detecting or recognizing individual objects (cf.…”