Volume 4 Issue 1
Authors: Jimmy Nagau; Jean-Luc Henry
Abstract: In automatic image analysis, when using natural images, it is important to select the elements in a scene that can lead to correct characterization. Extraneous elements can result in poor values in the parameter evaluation for recognition. Irrelevant elements also increase the processing time, because they can be used incorrectly by the algorithms for characterization or identification. The goal of this paper is to offer identification algorithms and characterizations that will allow effective recognition. We base our procedure on the photographer focus areas to identify those relevant elements in a scene. In this work, we propose a procedure to select the relevant shape from natural images. The treatment uses focal depth with edge detectors. The resulting points are combined with a region partition to obtain relevant shapes.
Keywords: Mathematical Morphology; Edge Detects; Focal Depth; Image Processing; Computer Vision
Fingerprints for Imposed Layers in Document Images Based on Huffman Code and Logical Layout Analysis
Authors: Surabhi Narayan; Sahana D Gowda
Abstract: A document is characterized by its layout and component structure. Document layout is due to the placement of the content components and document structure is due to the geometrical shape of the content components. Content components in a filled-in document image consist of general information foreground layer and vital information imposed layer. The foreground layer consists of printed text, logos, tables and lines that are identical for documents of the same class; the imposed layer of the document image consists of handwritten text, signatures and seals imposed on the document image that are unique to every document image. Processing filled-in document images for indexing, considering general information along with vital information is complex with the possibility of generating identical indexes due to large amount of general information suppressing fewer imposed layer vital information. In this paper, a novel technique was proposed to generate a unique code by formulating a logical layout of the imposed layer which was extracted from the filled-in document image using registration. The extracted imposed layer components were represented by centroids based on their spatial occupancy and the imposed layer was hierarchically decomposed into 16 equal quadrants. The Huffman tree generation algorithm was applied based on the number of centroids in a quadrant and with quadrant indices were assimilated to generate a unique code for the logical layout of the document image. In order to verify the applicability of this method, extensive experimentation were conducted on extracted imposed layers from application forms, student records, bank cheques and declaration forms.
Keywords: Imposed Layer; Centroids; Huffman Codes; Quad Decomposition; Logical Layout
Authors: Mohammed O. Assayony; Sabri A. Mahmoud
Abstract: In recent years, feature learning approaches have gained substantial interest and are successfully applied to challenging problems in facial recognition, visual object retrieval and classification, document image analysis and, recently, in handwriting recognition. In this paper, we present a feature learning framework for Arabic handwritten text recognition based on the Bag-of-Feature (BoF) paradigm. Utilizing the characteristics of handwritten text, we developed two novel versions of SIFT that are discriminative and computationally efficient with half the size of the original SIFT descriptors. To evaluate the quality of the features learned by the framework and the efficiency of the proposed versions of SIFT, we conducted extensive experimental work on two Arabic handwritten text datasets, viz. the non-touching Arabic Indian digit and Arabic sub-words datasets of CENPARMI Bank check database. Our framework achieves state-of-the-art accuracies on both datasets. The recognition performance and the computational efficiency are the result of utilizing the unique properties of the handwritten text.
Keywords: Feature Learning; Bag-of-Features; Arabic Handwriting Recognition; SIFT