Font Size: a A A

Research On The Fast Recognition Of Targets In Natural Scenes Based On EEG

Posted on:2015-12-13Degree:MasterType:Thesis
Country:ChinaCandidate:F YangFull Text:PDF
GTID:2208330431969159Subject:Computer technology
Abstract/Summary:PDF Full Text Request
Natural Scene Recognition is an important part of the cognitive process.We live in the nature scene, observe the world through the nature scene.Since Epstein found there is a specialized brain areas belong parahippocampal gyrus of human brain, this area is responsible for processing the scene information.Scene recognition in recent years has been one of the central issue of cognitive science.This paper reviews the process of the natural scene recognition.First,introduced the related concepts of cognitive, described the contents and mechanisms of cognitive research.Summarizes some prior models and methods for natural scene recognition, describes the current expansion of research about the natural scene.Based on the picture information, do research about the cognitive processes of animal and traffic tool in the natural scene and cognitive processes of animal and non-animal in the natural scene.In experiment1, we use animal picture and traffic tool picture as stimulate picture, all of this pictures shooting in natural scenes.All stimulate picture processing by SHINE Toolbox, eliminates the differences between low-lever attributes.Every picture conducted histogram matching, and has the same histogram.During the experiment, using Quest to control the contrast of the stimulates.Experiment1based on continuous flash suppression paradigm, collect ERP data by NeuroScan.By analyzing ERP data,we found that using continuous flash suppression under consciousness, subject has different recognitive performance between "animal" and "traffic tool" stimulate picture.In experiment2, we use different animal pictures and non-animal picture as stimulates picture, all of this pictures shooting in natural scenes.All stimulate picture processing by SHINE Toolbox, eliminates the differences between low-lever attributes.Every picture conducted histogram matching, and has the same histogram.During the experiment, using Quest to control the contrast of the stimulates.Experiment2based on continuous flash suppression paradigm, collect ERP data by NeuroScan.By analyzing ERP data,we found that using continuous flash suppression under consciousness,subject has different recognitive performance between "animal" and "non-animal" stimulate picture.
Keywords/Search Tags:image information, visual cognitive, natural scene, CFS, ERP
PDF Full Text Request
Related items