自然场景的识别及其理论模型  

The Recognition of Nature Scene and Its Models

作　　者：张序堃 孙雨生[1] 张智君[1] ZHANG Xu-kun;SUN Yu-sheng;ZHANG Zhi-jun(Department of Psychology and Behavioral Science,Zhejiang University,Hangzhou 310028,China)

机构地区：[1]浙江大学心理与行为科学系

出　　处：《应用心理学》2019年第3期205-217,共13页Chinese Journal of Applied Psychology

基　　金：国家自然科学基金(31371039)

摘　　要：人类能快速而高效地处理自然场景图片中的背景和客体信息。认知神经科学研究发现,大脑的旁海马位置区、枕叶位置区和扣带回的压后皮层等区域与场景识别息息相关。本文系统地总结和分析了场景识别领域的上述成果,并进一步介绍了其认知模型和计算模型。其中,前者阐述了前馈加工和反馈加工在场景识别过程中的作用和关系,而后者则总结了旨在理解场景识别过程的"算法"及其发展,并解释这些计算模型仍不能真正地"理解"场景的原因。Human beings can rapidly and accurately process the object information,as well as the context information,of the nature scene image.Numerous cognitive neuroscience studies have demonstrated that the parahippocampal place area(PPA),occipital place area(OPA) and the retrosplenial complex(RSC) play an important role in nature scene recognition.Based on these findings,we introduced a few models of nature scene recognition.Firstly,we summarized the cognitive models of nature scene recognition in which we illustrated the role of feedforward processing and feedback processing,as well as their interaction,during scene recognition.Secondly,we summarized the algorithm development of scene recognition and we also proposed some prospects to enrich the computational models so that these models could truly understand nature scene.

关 键 词：自然场景识别 前馈模型 预期 计算模型 

分 类 号：B842[哲学宗教—基础心理学]