Suppr超能文献

场景上下文在物体识别中的作用:时机与机制。

The role of scene context in object recognition: Timing and mechanisms.

作者信息

Gao Mingjie, Drewes Jan, Zhu Weina

机构信息

School of Information Science, Yunnan University, Kunming, China.

Institute of Brain and Psychological Sciences, Sichuan Normal University, Chengdu, China.

出版信息

J Vis. 2025 Aug 1;25(10):4. doi: 10.1167/jov.25.10.4.

Abstract

One remarkable aspect of human perception is the ability to extract the meaning of a target within a complex scene quickly. Consistent scenes between targets and backgrounds enhance visual understanding, yet the precise timing and underlying mechanisms of this effect remain unclear. To address this, two experiments were conducted exploring how object category, scene orientation, and scene consistency influence object recognition under conscious (Experiment 1) and unconscious (Experiment 2) conditions. Both experiments revealed an animate advantage, facilitation by scene consistency, and orientation-specific enhancements in object recognition. Performance was improved when both the object and the scene were upright, and furniture targets were more affected by scene consistency than animal targets. Specifically, under the conscious condition, the animate advantage was observed only for inconsistent scenes, whereas under the unconscious condition, the animate advantage was not influenced by scene consistency. Interestingly, in the unconscious state, the effects of target category and background orientation depended on scene consistency, with animals consistent with the background and furniture inconsistent with the background both influenced by orientation. These results suggest that scene context influences object recognition in the early stage of visual processing, and furniture recognition is more sensitive to contextual regularities than animal recognition.

摘要

人类感知的一个显著特点是能够在复杂场景中迅速提取目标的意义。目标与背景之间的一致场景能增强视觉理解,但这种效应的确切时间和潜在机制仍不清楚。为解决这一问题,进行了两项实验,探究在有意识(实验1)和无意识(实验2)条件下,物体类别、场景方向和场景一致性如何影响物体识别。两项实验均揭示了有生命物体的优势、场景一致性的促进作用以及物体识别中特定方向的增强效应。当物体和场景都呈直立状态时,表现会得到改善,并且家具目标比动物目标更容易受到场景一致性的影响。具体而言,在有意识条件下,有生命物体的优势仅在不一致场景中观察到,而在无意识条件下,有生命物体的优势不受场景一致性的影响。有趣的是,在无意识状态下,目标类别和背景方向的效应取决于场景一致性,与背景一致的动物和与背景不一致的家具都受到方向的影响。这些结果表明,场景背景在视觉处理的早期阶段影响物体识别,并且家具识别比动物识别对上下文规律更敏感。

https://cdn.ncbi.nlm.nih.gov/pmc/blobs/89b3/12347156/8dd8eed2a913/jovi-25-10-4-f001.jpg

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验