基于兴趣目标的图像检索

当前图像检索算法通常针对整体图像提取特征以完成检索任务.然而,在很多情况下用户只会关注图像的一部分,即他们的兴趣目标.此时,从整体图像提取的特征一部分是有效的,另一部分则是无效的且会对检索过程带来消极影响.为此,本文提出基于兴趣目标的图像检索方案,并借助于现有的显著性检测、图像分割、特征提取等技术实现一款有效的图像检索算法.首先采用HS （Hierarchical Saliency,分层显著性）检测算法分析用户的兴趣目标并应用SC （Saliency-based Image Cut,基于显著性的图像分割）算法将其分割,然后针对兴趣目标提取HSV （Hue、Saturation、Value,色调、饱和度、明度）颜色特征、SIFT （Scale Invariant Feature Transform,尺度不变特征变换）局部特征和CNN （Convolutional Neural Network,卷积神经网络）语义特征,最后计算其与数据库图像的相似度并根据相似度排序返回检索结果.仿真实验结果表明,本文算法在解决"这是什么东西"这类图像检索任务时明显优于现有的图像检索算法.

Image Retrieval Based on Interested Objects

The current image retrieval algorithms usually extract features from the whole input image to conduct retrieval tasks.However,in many cases users focus on only a part of the image,i.e.object-of-interest.As a result,the features extracted from the image are partially effective.In other words,some of the features are ineffective and might have a negative impact on the retrieval process.To overcome this difficulty,an image retrieval scheme based on object-of-interest is proposed.By incorporating this retrieval scheme with the existing techniques in saliency detection,image segmentation,and feature extraction,an effective image retrieval algorithm is coded.First,the hierarchical saliency (HS) detection algorithm is adopted to analyze the user's object-of-interest,and the saliency-based image cut (SC) algorithm is employed to segment it from the input image.Then,we extract the hue,saturation,value (HSV) color features,the scale invariant feature transform (SIFT) local features and the convolutional neural network (CNN) semantic features of the object-of-interest.Finally,the similarity of object-of-interest between a query image and every database image is computed and the retrieval result is sorted accordingly.Simulation experimental results show that,when being used to cope with a retrieval task like "what is this",the proposed algorithm is significantly better than the current image retrieval algorithms.