Research And Application Of Feature-based Document Image Retrieval

As an important part of image retrieval, Document image retrieval has wide applications on digital library, AO (Automatic Office); etc. Document image retrieval aims at finding out one or more images which meet the requirements from the document image database. Ordinary document image retrieval algorithms can be classed into two classes, document layout reconstruction based image retrieval and image level feature based method. In the feature based image retrieval, it is difficult to represent document images using visual features such as color and texture etc.. And it is difficult too to extract features from the whole to represent the content of document images. Therefore, one of the key in the document image retrieval based on feature is how to extract the features and how to calculate the characteristics of similarity.Based on the analysis of the existing feature-based document image retrieval methods'drawbacks and virtues, a novel retrieval method is proposed .It falls in with the basic procedures of feature-based document image retrieval methods and takes advantage of sub-block method in content-based image retrieval methods. Firstly, preprocessing of the document image is done. It includes noise-removing and skew detecting .Using Median filter to remove the salt and pepper noise is efficient. As the first step of skew detecting, the binarization utilizes the Bernsen algorithm. Then a quick tilt testing method is used to detect the tilt of the document images and a fast image rotation algorithm is used for rotation correction. After the preprocessing, the length and width of effective area, density feature, and SIFT feature are then defined and extracted from the whole image. The next step is to segment the image into text areas and non-text areas, an efficient segmentation using the maximum gradient (MGD) is presented. Based on the text area, local features such as the distances between components (the lengths of gaps), heights of components and widths of components are extracted. And global features such as the number of connected component, the number of cavities, average height of the components, average width of the components, average distance between components and paragraph feature are extracted. As for the non-text areas, the key block feature is extracted. The SIFT feature is invariant to scalability, shift and distortion. This renders it robust to deformation of the document image. Features extracted from text area are low level features, and could characterize the document in deep degree. Density feature as well as key block feature is proved to be efficient in describing the document image. In other words, the features defined and extracted in this paper are efficient. They include not only global features but also local features. And low level features as well as high level ones are all covered. Hence, the combination of these features gives sufficient representation of the image. These features are grouped into 3 feature vectors according to their lengths and properties. As a High-dimensional index structure appropriate for indexing very high dimensional data, Clustering Pyramid-tree is used in the paper to organize these features extracted from the database of document images. Each feature vector is used to create an Clustering Pyramid-tree respectively. By inquiring each Clustering Pyramid-tree, three set is then got, and the candidate set is then created by the union of them. After updating the weight of the candidate set, the final answers are given according to the weight. The retrieval method proposed is adaptive to handwritten or printed document images.The experiments are performed on a database which consist of 4300 text dominated images and images mixed with texts, pictures and tables respectively in order to give an experimental verification of the adaptation of these features. The results show that the proposed method has good performance and robustness, it is a practical method.