【摘要】 对图像进行准确标注是提高图像检索率的有效手段,针对目前图像标注的相关研究存在的问题,本研究提出了一种基于神经网络的图像标注模型,此模型由一个自适应分类网络和一个非线性相关网络组成。模型进行图像标注分为两个阶段:第一阶段,从输入图像分割出的不同区域提取特征信息传送到自适应分类网络,以产生分类标签;第二阶段,非线性相关网络以通过训练图像学习的关键词的相关性为依据来细化分类结果。为验证模型标注的准确性,选取LabelMe和Caltech-101图像数据库中的图像进行相关实验,结果表明本研究提出的模型提高了图像标注的准确率。更多还原

【Abstract】 In this study,an image annotation model based on neural network was proposed aiming to improve the image retrieval.This model consisted of an adaptive classification network and a nonlinear correlation network.The image annotation with the model was divided into two phases.In the first stage,the features were extracted from the different segmentation regions of the input image,and sent to the adaptive classification network to produce a classification label.In the second stage,the nonlinear network refined the classification results depending on the correlation of the keywords of training image.Related experiments were carried out based on LabelMe and Caltech-101 image database.The experimental results showed that the proposed model can improve the accuracy of image annotation.更多还原