Document image retrieval based on texture features and similarity fusion

Abstract

In this paper we investigate the usefulness of two different texture features along with classification fusion for document image retrieval. A local binary texture method, as a statistical approach, and a wavelet analysis technique, as a transform-based approach, are used for feature extraction and two feature vectors are obtained for every document image. The similarity distances between each of the two feature vectors extracted for a given query and the feature vectors extracted from the document images in the training step are computed separately. In order to use the properties of both features, a classifier fusion technique is then employed using a weighted average fusion of distance measures obtained in relation to each feature vector. The document images are finally ranked based on the greatest visual similarity to the query obtained from the fusion similarity measures. The Media Team Document Database, which provides a great variety of page layouts and contents, is considered for evaluating the proposed method. The results obtained from the experiments demonstrate a correct document retrieval of 65.4% and 91.8% in the Top-1 and Top-10 ranked document list, respectively.

Document Image Retrieval Based on Visual Saliency Maps

“…To have a fast and effective retrieval system, a fast measuring distance is necessary. In the previous studies, the k-nearest neighbor (k-NN) method has usually been used to find the similarity distance between a query image and trained document images [10,11]. Since we have a large dataset, locality-sensitive hashing (LSH) [13] is taken into account in this work.…”

“…4. The methods that used a combination of the Gist and wavelet transform features [23], as well as the LBP and wavelet transform features [11] for document image IV. CONCLUSION In the present study, an efficient document image retrieval method based on human visual system/attention is presented.…”

“…Texture features are generally fast and simple to compute and they are computationally suitable for large volumes of data. Many researchers from the document image analysis (DIA) community have also used various texture features in different stages of DIA systems, including segmentation, layout analysis, recognition [11,12], etc. Furthermore, texture features have been used to discriminate textual and nontextual areas, and to differentiate between various types of text fonts and graphics in document images [13].…”

Fast local binary pattern: Application to document image retrieval

“…Various feature extraction techniques have also been developed for document image analysis and retrieval [8][9][10]. The advantages and beneficial uses of texture features for document image retrieval have further been discussed in [11,12].…”

“…The effectiveness of texture features has been explored using document image databases. Likewise in [12], the LBP and discrete wavelet transform (DWT) features have been extracted from each document image to characterize document images. A score level fusion strategy has then been proposed to obtain the final retrieval results in the proposed document image retrieval system [12].…”