Legal case summarization is an important problem, and several domain-specific summarization algorithms have been applied for this task. These algorithms generally use domain-specific legal dictionaries to estimate the importance of sentences. However, none of the popular summarization algorithms use document-specific catchphrases, which provide a unique amalgamation of domain-specific and document-specific information. In this work, we assess the performance of two legal document summarization algorithms, when two different types of catchphrases are incorporated in the summarization process. Our experiments confirm that both the summarization algorithms show improvement across all performance metrics, with the incorporation of document-specific catchphrases.

In this study, a robust algorithm for dewarping of camera‐captured document images, mainly in Bangla script, is proposed. The algorithm can handle various types of warped document images and they are generated due to different types of document surfaces (convex, concave or multi‐folded). The proposed algorithm is independent of font type, font size, font style and camera view angle. After initial preprocessing, the method first demarcates the text lines present in the document image. Then, the headline ( shirorekha ) position of each text line is estimated. Based on the headline position and shape, each text line is dewarped. If the document is highly warped, distorted text (e.g. thinner and shorter characters) is generated after dewarping. Special care has been taken to minimise this distortion based on most undistorted character information. Exhaustive testing shows the robustness and shape improvement of the proposed algorithm. Finally, for shape quality evaluation, some new measures are defined.

In this paper, we propose a novel multiple pyramids based image inpainting method using local patch statistics and geometric feature-based sparse representation to maintain texture consistency and structure coherence. First, we approximate each patch in the target region (region to be inpainted) by statistically dominant local candidate patches to preserve local consistency. Then each approximated patch is refined by a sparse representation of candidate patches based on local steering kernel (LSK) feature to retain texture quality. We also propose a multiple pyramids based approach to generate several inpainted versions of the input image, one for each of the pyramids. Finally, we combine the inpainted images by gradient-based weighted average to produce the final inpainted image. This approach helps to maintain structure coherence and to remove artifacts which may appear in the inpainted images due to different initial scales of the individual pyramids. The proposed method is tested on a wide range of natural images for scratch and blob/object removal. We have presented both quantitative and qualitative comparison with the existing methods to demonstrate the superiority of the proposed method.

This paper presents a Markov random field (MRF)-based image inpainting algorithm using patch selection from groups of similar patches and optimal patch assignment through joint patch refinement. In patch selection, a novel group formation strategy based on subspace clustering is introduced to search the candidate patches in relevant source region only. This improves patch searching in terms of both quality and time. We also propose an efficient patch refinement scheme using higher order singular value decomposition to capture underlying pattern among the candidate patches. This eliminates random variation and unwanted artifacts as well. Finally, a weight term is computed, based on the refined patches and is incorporated in the objective function of the MRF model to improve the optimal patch assignment. Experimental results on a large number of natural images and comparison with well-known existing methods demonstrate the efficacy and superiority of the proposed method.

PDF format of scanned document images is not searchable. OCR tries to remedy this adversity by converting document images into editable and searchable data, but it has its own limitations in presence of equations - both mathematical and chemical. OCR system for mathematical equation is already a major research area and has provided successful result. However, chemical equation segmentation has been a less ventured road. In this paper, we present a novel method for automated generation of searchable PDF format of segmented chemical equations from scanned document images by performing chemical symbol recognition and auto-correction of OCR output. We use existing OCR system, pattern recognition technique, contextual data analysis and a standard LaTeX package to generate the chemical equation in searchable PDF format. The effectiveness of the proposed method is verified through exhaustive testing on 234 document images.

This paper presents a sparse representation based image inpainting method using local patch analysis and geometric structure based feature extraction. In local patch analysis, we approximate the target region by weighted average of some local patches which are frequently occurred within a neighborhood. Local patch statistics is applied to find the most relevant neighbors for each target patch. Further we extract local steering kernel (LSK) based feature to preserve geometric structure and texture sharpness in the target region. The advantage of non local self similarity as redundancy of similar patches in natural images is introduced to find the candidate patches from the whole source region. Based on these local and non local prior information we propose a sparse representation framework for image inpainting. Our proposed method is tested on wide range of natural images. The experimental results show the superiority of the proposed method compared to some of the previous approaches.

Paper maps need be preserved for their historical importance. These maps are characterized by the presence of texts of different fonts, sizes and orientations on a complex background. These texts are cluttered with criss-cross lines representing boundaries, rivers, roads and railway lines, and many other objects. So, for scanned map image segmentation, representation or vectorization is difficult and make automatic digitization challenging. This paper presents a framework for automatic region segmentation, vectorization and reconstruction of regions within a map. Each region is segmented by analysing histogram of hue, then the contours are extracted, and at last each contour is split into several smaller line/curve segments. Bezier curve fitting is used for representation of smaller curves and finally the whole contour is reconstructed from a minimum number of known points. This way all regions within a map are reconstructed. The approach is tested on a map image dataset of moderate size and the results are encouraging.

An electrical or an electronic circuit often contains special combinations of circuit symbols in the form of sub-circuits. Identification and analysis of these sub-circuits can substantially simplify the underlying topology of a circuit. This paper explains our maiden attempt towards topological simplification of a circuit by identifying two important classes of sub-circuits—one formed by symbols connected in series and another by symbols connected in parallel. Although a sub-circuit is usually formed with the same type of circuit symbols, our method can handle any combination, irrespective of the symbol types. The method, in general, is based on a novel histogram analysis, mathematical morphology, and geometric features during the symbol identification phase, and a set of adjacency lists representing the connectivity matrix during the analysis phase. The idea may be extended to match topologically similar but spatially different circuits using graph isomorphism. It may also be used for vectorization of circuit drawings utilizing the information on the segmented circuit elements and their connectivity matrices. We have tested the proposed method on a dataset containing 83 scanned images of a variety of electronic and electrical drawings. Some of the results are presented here to demonstrate its efficacy and robustness.

This work presents a binarization technique of map document images. It exploits an amalgam of global and local threshold approaches best suited for binarization of document images with complex background and overlapping objects in the foreground like maps. The proposed approach uses Distance Transform (DT) and Adaptive threshold. Initially a rough estimate of the map background is done using Distance Transform (DT). This is followed by an adaptive threshold operation to extract the foreground. The efficacy and accuracy of the proposed technique are also compared, using ICDAR 2009 -- DIBCO (Document Image Binarization Contest) evaluation parameters, with other algorithms already reported in the literature.

Text, graphics and half-tones are the major constituents of any document page. While half-tone can be characterised by its inherent intensity variation, text and graphics share common characteristics except difference in spatial distribution. The success of document image analysis systems depends on the proper segmentation. The success of document image analysis systems depends on the proper segmentation of text and graphics as text is further subdivided into other classes such as heading, table and math-zones. Segmentation of graphics is essential for better OCR performance and vectorization in computer vision applications. Graphics segmentation from text is particularly difficult in the context of graphics made of small components (dashed or dotted lines etc.) which have many features similar to texts. Here we propose a robust technique for segmenting all sorts of graphics and texts in any orientation from document pages.

In this paper we propose a fully automatic hierarchical method for identification of forms using global as well as local features. Moments of certain orders are considered as global shape features and are utilised to reduce the search space by selecting a subset of forms present in the database. The type of the candidate form is then identified within this subset through detail analysis using local geometrical and topological features. The candidate form is then segmented to extract the user-filled information.

With an aim to high-level understanding of the mathematicalcontents in a document image the requirement ofmath-zone extraction and recognition technique is obvious.In this paper we present fully auotmatic segmentation ofdisplayed-math zones from the document image, using onlythe spatial layout information of math-formulas and equations,so as to help commercial OCR systems which cannotdiscern math-zones and also for the identification and arrangementof math symbols by others.

With an aim to extract the structural information from the table of contents (TOC) to help develop a digital document library, the requirement of identifying/segmenting the TOC page is obvious. The objective to create a digital document library is to provide a non-labour intensive, cheap and flexible way of storing, representing and managing the paper document in electronic form to facilitate indexing, viewing, printing and extracting the intended portions. Information from the TOC pages is to be extracted for use in a document database for effective retrieval of the required pages. We present a fully automatic identification and segmentation of a table of contents (TOC) page from a scanned document.