Search

Digitization offers a solution to the challenges associated with managing and retrieving paper-based documents. However, these paper-based documents must be converted into a format that digital machines can comprehend, as they primarily understand alphanumeric text. This transformation is achieved through Optical Character Recognition (OCR), a technology that converts scanned image documents into a format that machines can process. A novel top-down character segmentation approach has been proposed in this work, involving multiple stages. Our approach began by isolating lines from handwritten documents and using these lines to segment words and characters. To further enhance the character segmentation, a Raster Scanning object detection technique is employed to isolate individual characters within words. Thus, the character segmentation results are integrated from the results of the vertical projection and raster scanning. Recognizing the significance of advancing digitization of handwritten documents, we have chosen to focus on the regional languages of Assam and Andhra Pradesh due to their historical and cultural importance in India's linguistic diversity. So, we have collected datasets of handwritten texts in Assamese and Telugu languages due to their unavailability in the desired form. Our approach achieved an average segmentation accuracy of 93.61%, 85.96%, and 88.74% for lines, words, and characters for both languages. The key motivation behind opting for a top-down approach is two-fold: firstly, it enhances the accuracy of character recognition, and secondly, it holds the potential for future use in language/script identification through the utilization of segmented lines and words. [ABSTRACT FROM AUTHOR]

This paper presents an efficient architecture for two-dimensional (2-D) adaptive FIR filter architecture using the distributed arithmetic (DA) algorithm. DA-based filter architectures essentially require look-up tables (LUT). In the proposed filter architecture, RAM- or ROM-based LUT is replaced by adders- and logic gates-based structure that generates the LUT value corresponding to the input. Therefore, the MAC unit requires fewer logic gates and adders in DA-based realization. In addition, the memory sharing concept in architecture reduces the delay elements. Moreover, the complexity of the LUT hardware of higher-order filters is reduced by parallelly dividing the internal MAC block for the DA decomposition which offers a higher degree of modularity and parallelism in the proposed architecture. Further, 2-D delayed LMS algorithm is used for the updation of the filter coefficient weights. Furthermore, two-stage pipelining is used to reduce the critical path of the architecture and it also makes critical path delay independent of the order of the filter. ASIC synthesis results reveal the advantages of the proposed structure by reducing the area, power, ADP and EDP by 54%, 48.19%, 55% and 49%, respectively, as compared with the existing architecture for filter size 8 × 8. [ABSTRACT FROM AUTHOR]

The special features of the implementation and the main sources of measurement uncertainty of the power of an ultrasound beam in water are examined by the method of plane scanning of the cross-section of a beam. The methods of evaluating the effect of deviations from the conditions of a plane wave and a far field were refined. Methods of decreasing the laboriousness of measuring beams with rotational symmetry are substantiated. The results are presented for the comparison of the values of the power of ultrasound emission of a round piston emitter, obtained by the plane scanning method and by the reference method of equilibration of emission force. The sufficiently good concurrence of the results provides evidence of the potential possibility of increasing the precision of hydrophone calibrations by the methods of plane (raster or diametral) scanning. [ABSTRACT FROM AUTHOR]

The resolution and accuracy of terahertz (THz) imaging depend on the scanning distance between raster pixels. Recently, numerous studies have focused on increasing scanning speed and accuracy. The current study presented a high-speed scanning method without dwell time during scanning. A compensation factor for the velocity in a motorized x-y linear stage with a scanning sample was introduced to maintain the high accuracy of image scanning with the proposed method. The THz imaging system consisted of fiber-coupled pulsed photoconductive antenna and motorized x-y linear stage based on THz-time-domain spectroscopy. A validation specimen was fabricated by using copper tape and acrylic plate to compare the image scanning performance of various methods. Reflection mode scanning with 90° incident angles was conducted on the validation sample. The performance of the proposed scanning method without dwell time was evaluated and compared with that of other methods in terms of scanning speed, which is defined as the number of pixels per second; and geometric matching ratio, which uses modal assurance criterion (MAC) to quantify scanning accuracy. Results confirmed that the scanning speed of the proposed method is 328 % faster than that of conventional scanning methods and that the MAC value of the geometric matching ratio is more than 0.994 without any additional equipment installation. [ABSTRACT FROM AUTHOR]

This paper presents a generalized formulation of 2-D IIR filters using distributed arithmetic (DA) techniques. Based on the DA formulation, two efficient structures for 2-D IIR filters are proposed. Hardware-based look-up table (HLUT) is used in the internal blocks, so the proposed structures are reconfigurable. A novel approach of HLUT sharing, among the various internal blocks of structure, is used to reduce the requirement of adders and memory elements. For higher-order 2-D IIR filter, the complexity of HLUT is reduced by dividing the internal block of 2-D IIR filter into parallel and small block for DA decomposition. Such decomposition for higher-order 2-D IIR filters offers high degree of modularity, parallelism and regularity in building blocks, thereby achieving easier hardware and software implementation. In order to reduce combinational delay in the critical path, pipelining is used in the structures. Since proposed structures are multiplier-less and require lesser number of delays and adders, a significant improvement in chip area, power consumption and throughput can be obtained. Finally with the help of ASIC synthesis results, a comparative analysis is made and the results show that for the filter order 15, the proposed structures offer 69 and 59% reduction in ADP and 79 and 76% reduction in ADP than the earlier reported results for separable and non-separable structures, respectively. [ABSTRACT FROM AUTHOR]

Possible ways of improving laser image generators based on the circular scanning principle are analyzed. The most important sources of errors induced by the system of radial displacements and by the rotation unit are studied. Principles of decreasing the main components of instability of operation of image generators with circular scanning are described. Owing to the use of some principles, the prototype of such a generator can be put onto the second place in the worldwide rating of systems used by various companies to form structures providing high-accuracy measurements of angles (scales, limbs, rasters, etc.). [ABSTRACT FROM AUTHOR]

An algorithm for Cartesian trajectory generation by redundant robots in environments with obstacles is presented. The algorithm combines a raster scanning technique, genetic algorithms and functions for interpolation in the joint coordinates space in order to approximate a desired Cartesian curve by the robot's hand tip under maximum allowed position deviation. A raster scanning technique determines a minimal set of knot points on the desired curve in order to generate a Cartesian trajectory with bounded position approximation error. Genetic algorithms are used to determine an acceptable robot configuration under obstacle avoidance constraints corresponding to a knot point. Robot motion between two successive knot points is finally achieved using well known interpolation techniques in the joint coordinates space. The proposed algorithm is analyzed and its performance is demonstrated through simulated experiments carried out on planar redundant robots. [ABSTRACT FROM AUTHOR]