本文提出了一种基于关键词的中文文档图像检索方法,能在不经OCR(Optical Character Recognition)识别的情况下,直接利用中文字符的图像特征进行关键词检索。首先将文档图像分割成单个中文字符图像,接着对字符图像进行汉字笔画的特征数...本文提出了一种基于关键词的中文文档图像检索方法,能在不经OCR(Optical Character Recognition)识别的情况下,直接利用中文字符的图像特征进行关键词检索。首先将文档图像分割成单个中文字符图像,接着对字符图像进行汉字笔画的特征数据提取,然后在特征数据间进行基于WMHD(Weighted Modified Hausdorff Dis-tance)的相似性测量。该方法不受字号的影响,也有一定的抗字体能力,实验证明其具有较高的检索效果。展开更多
The development of document image databases is becoming a challenge for document image retrieval tech-niques.Traditional layout-reconstructed-based methods rely on high quality document images as well as an optical ch...The development of document image databases is becoming a challenge for document image retrieval tech-niques.Traditional layout-reconstructed-based methods rely on high quality document images as well as an optical char-acter recognition(OCR)precision,and can only deal with several widely used languages.The complexity of document layouts greatly hinders layout analysis-based approaches.This paper describes a multi-density feature based algorithm for binary document images,which is independent of OCR or layout analyses.The text area was extracted after prepro-cessing such as skew correction and marginal noise removal.Then the aspect ratio and multi-density features were extract-ed from the text area to select the best candidates from the document image database.Experimental results show that this approach is simple with loss rates less than 3%and can efficiently analyze images with different resolutions and dif-ferent input systems.The system is also robust to noise due to its notes and complex layouts,etc.展开更多
针对传统近重复文本图像检索方法需人工事先确定近重复文本图像之间存在的变换类型,易受到人主观性影响这一问题,提出一个面向近重复文本图像检索的三分支孪生网络,能自动学习图像之间存在的各种变换。该网络输入为三元组,包括查询图像...针对传统近重复文本图像检索方法需人工事先确定近重复文本图像之间存在的变换类型,易受到人主观性影响这一问题,提出一个面向近重复文本图像检索的三分支孪生网络,能自动学习图像之间存在的各种变换。该网络输入为三元组,包括查询图像、查询图像的近重复图像以及其非近重复图像,训练时采用三元损失使得查询图像和近重复图像之间的距离小于查询图像与非近重复图像之间的距离。提出的方法在两个数据集上的mAP(mean average precision)分别达到98.76%和96.50%,优于目前已有方法。展开更多
This paper investigated approaches to supporting effective and efficient retrieval of image based on principle component analysis. First, it extracted the image content, texture and color. Gabor wavelet transforms wer...This paper investigated approaches to supporting effective and efficient retrieval of image based on principle component analysis. First, it extracted the image content, texture and color. Gabor wavelet transforms were used to extract texture feature of the image and the average color was used to extract the color features. The principle component of the feature vector of image can be constructed. Content based image retrieval was performed by comparing the feature vector of the query image with the projection feature vector of the image database on the principle component space of the query image. By this technique, it can reduce the dimensionality of feature vector, which in turn reduce the searching time.展开更多
基金supported by the National Natural Science Foundation of China(Grant No.60472028)the Specialized Research Fund for the Doctoral Program of Higher Education(No.20040003015).
文摘The development of document image databases is becoming a challenge for document image retrieval tech-niques.Traditional layout-reconstructed-based methods rely on high quality document images as well as an optical char-acter recognition(OCR)precision,and can only deal with several widely used languages.The complexity of document layouts greatly hinders layout analysis-based approaches.This paper describes a multi-density feature based algorithm for binary document images,which is independent of OCR or layout analyses.The text area was extracted after prepro-cessing such as skew correction and marginal noise removal.Then the aspect ratio and multi-density features were extract-ed from the text area to select the best candidates from the document image database.Experimental results show that this approach is simple with loss rates less than 3%and can efficiently analyze images with different resolutions and dif-ferent input systems.The system is also robust to noise due to its notes and complex layouts,etc.
文摘针对传统近重复文本图像检索方法需人工事先确定近重复文本图像之间存在的变换类型,易受到人主观性影响这一问题,提出一个面向近重复文本图像检索的三分支孪生网络,能自动学习图像之间存在的各种变换。该网络输入为三元组,包括查询图像、查询图像的近重复图像以及其非近重复图像,训练时采用三元损失使得查询图像和近重复图像之间的距离小于查询图像与非近重复图像之间的距离。提出的方法在两个数据集上的mAP(mean average precision)分别达到98.76%和96.50%,优于目前已有方法。
文摘This paper investigated approaches to supporting effective and efficient retrieval of image based on principle component analysis. First, it extracted the image content, texture and color. Gabor wavelet transforms were used to extract texture feature of the image and the average color was used to extract the color features. The principle component of the feature vector of image can be constructed. Content based image retrieval was performed by comparing the feature vector of the query image with the projection feature vector of the image database on the principle component space of the query image. By this technique, it can reduce the dimensionality of feature vector, which in turn reduce the searching time.
