Computer Recognition Research Articles

Multi-scale topographic analysis of Merbabu and Merapi volcanoes using wavelet decomposition

With the increasing availability of data, geo-sciences have experienced deep changes in handling and processing it. One of the presently explored research directions concerns the systematic decomposition and understanding of topographical features, without the subjective interactions of humans. This can eventually lead to fully automated algorithms for topographic analysis and understanding. This paper aims at being a contribution to this broad research area for the specific cases of stratovolcanoes, whose general geometry are very similar to a perfect cone. More specifically, this paper addresses two issues: (1) is it possible to separate erosion features (local variations) from structural features (large variations) on stratovolcanoes, through mathematical expression; (2) can information on volcanic activity—intensity, age, etc.—be retrieved from a topographic analysis? The study has been conducted from two volcanoes in Central Java (Indonesia): the Merapi and the Merbabu. The DEM of these volcanoes has been sampled using concentric circles with a radius ranging from 500 to 5,000 m (horizontal distance) to the summit. The data conversion and sampling was performed in ArcMap®, while the data analysis was carried out with Matlab®, using Discrete Meyer wavelet decomposition. Results provide an insight on large-scale topographic variations (long-wave wavelet) that have been separated from rapidly varying topographic features such as lahar channels (short-wave wavelets). Observations proved that flanks where the most recent volcanic activity occur—like at Merapi Volcano on the S-SE flank—present a very low variability of long-wave variations, whereas short-wave variations are important. The author argues that this feature is due to highly erosive lahars that dig the valleys combined with a recent production of material and volcanic growth keeping the overall structure regular. Flanks with lesser activity are characterized at the two volcanoes by important long-wave variations—most certainly due to long-term differential erosion—and different level of short-waves variations. Comparing the two volcanoes, results show that the valleys of Merapi and Merbabu volcanoes are deeply incised, indicating recent periods of high activity, with reworkable material eroded by lahars and other channels deepening processes. The topography of the summit area of Merapi Volcano is smoother than at Merbabu Volcano, where deep erosion features extend up to the summit area. This difference is most certainly due by the material production at the summit of Merapi Volcano. Developing such classification is important for automated mapping and computer recognition of volcanic past activities and their impacts on landscapes. It is the based for the development of decision trees that assist computer assisted and automated computer vision.

Human and computer recognition of regional accents and ethnic groups from British English speech

The paralinguistic information in a speech signal includes clues to the geographical and social background of the speaker. This paper is concerned with automatic extraction of this information from a short segment of speech. A state-of-the-art language identification (LID) system is applied to the problems of regional accent recognition for British English, and ethnic group recognition within a particular accent. We compare the results with human performance and, for accent recognition, the ‘text dependent’ ACCDIST accent recognition measure. For the 14 regional accents of British English in the ABI-1 corpus (good quality read speech), our LID system achieves a recognition accuracy of 89.6%, compared with 95.18% for our best ACCDIST-based system and 58.24% for human listeners. The “Voices across Birmingham” corpus contains significant amounts of telephone conversational speech for the two largest ethnic groups in the city of Birmingham (UK), namely the ‘Asian’ and ‘White’ communities. Our LID system distinguishes between these two groups with an accuracy of 96.51% compared with 90.24% for human listeners. Although direct comparison is difficult, it seems that our LID system performs much better on the standard 12 class NIST 2003 Language Recognition Evaluation task or the two class ethnic group recognition task than on the 14 class regional accent recognition task. We conclude that automatic accent recognition is a challenging task for speech technology, and speculate that the use of natural conversational speech may be advantageous for these types of paralinguistic task.

Parallel processing for stepwise generalisation method on multi-core PC cluster

An approximate query, which is an approximate pattern matching in sequence databases, is one of the most important techniques for many different areas, such as computational biology, text mining, web intelligence and pattern recognition; it returns many similar sub-sequences. In this paper, we refer to a set of such similar sub-sequences as a mismatch cluster. To support users who execute an approximate query on a sequence database to find the regularities of approximate patterns that similar to the query pattern, we have developed the stepwise generalisation method that extracts a reduced expression, called a minimum generalised set, from a mismatch cluster. This paper proposes a novel parallelisation model with a hierarchical task pool for the parallel processing of the stepwise generalisation method on a multi-core PC cluster. To manage tasks efficiently on multi-core CPUs, the proposed model uses the hierarchical task pool and an efficient hierarchical dynamic load balancing technique. We evaluate the proposed method using real protein sequences on an actual multi-core PC cluster. Experimental results confirm that the proposed method performs well on multi-core CPUs and on a multi-core PC cluster.

Regularized Discriminant Analysis for Holistic Human Activity Recognition

A holistic or appearance-based eigenfeature regularization methodology based on a three-parameter eigenmodel improves computer recognition of human activity.

New Electric Tomography Scheme and Algorithms for Fast Recognition of Compound Structures

ABSTRACTNecessity of recognition of compound structures appears in different applications. Electrical tomography is one of perspective approach for investigation of material structures. A new special measurement scheme in electric tomography is proposed for the domain as a circle. This scheme opens possibility to construct electric tomography apparatus with faster computer recognition of structure of compounds materials, using General Ray algorithms. Computer simulation of this scheme is realised with MATLAB software and justified by numerical experiments on examples, related with electrical tomography.

Automatic Bubble Detection in Cardiac Video Imaging

Bubble recognition is a challenging problem in a broad range from mechanics to medicine. These gas-filled structures whose pattern and morphology alter in their surrounding environment would be counted either manually or with computational recognition procedures. In cardiology, user dependent bubble detection and temporal counting in videos require special trainings and experience due to ultra fast movement, inherent noise and video quality. In this study, we propose an efficient recognition routine to increase the objectivity of emboli detection. Firstly, we started to compare five different methods on two synthetic data sets emulating cardiac chamber environment with increasing speckle noise levels. Secondly, real echocardiographic video records were segmented by variational active contours and Left Atria (LA) were extracted. Finally, three successful methods in simulation were applied to LAs in order to reveal candidate bubbles on video frames. Our detection rate of proposed method was 95.7% and the others were 86.2% and 88.3%. We conclude that our approach would be useful in long lasting video processing and would be applied in other disciplines.

A general computational recognition primed decision model with multi-agent rescue simulation benchmark

Analytical decision making strategies rely on weighing pros and cons of multiple options in an unbounded rationality manner. Contrary to these strategies, recognition primed decision (RPD) model which is a primary naturalistic decision making (NDM) approach assumes that experienced and professional decision makers when encounter problems in real operating conditions are able to use their previous experiences and trainings in order to diagnose the problem, recall the appropriate solution, evaluate it mentally, and implement it to handle the problem in a satisficing manner. In this paper, a computational form of RPD, now called C-RPD, is presented. Unified Modeling Language was used as a modeling language to represent the proposed C-RPD model in order to make the implementation easy and obvious. To execute the model, RoboCup Rescue agent simulation environment, which is one of the best and the most famous complex and multi-agent large-scale environments, was selected. The environment simulates the incidence of fire and earthquakes in urban areas where it is the duty of the police forces, firefighters and ambulance teams to control the crisis. Firefighters of SOS team are first modeled and implemented by utilizing C-RPD and then the system is trained using an expert’s experience. There are two evaluations. To find out the convergence of different versions developed during experience adding, some of the developed versions are chosen and evaluated on seven maps. Results show performance improvements. The SOS team ranked first in an official world championship and three official open tournaments.

Computational recognition of gene boundaries

Computational recognition of gene boundaries

Mass-Spring-Damper Model for Offline Handwritten Character Distortion Analysis

Among the various aspects of offline handwritten character patterns, it is the great variety of writing styles and variations that renders the task of computer recognition very hard. The immense variety of character shape has been recognized but rarely studied during the past decades of numerous research efforts. This paper tries to address the problem of measuring image distortions and handwritten character patterns with respect to reference patterns. This work is based on mass-spring mesh model with the introduction of simulated electric charge as a source of the external force that can aid decoding the shape distortion. Given an input image and a reference image, the charge is defined, and then the relaxation procedure goes to find the optimum configuration of shape or patterns of least potential. The relaxation process is based on the fourth order Runge-Kutta algorithm, well-known for numerical integration. The proposed method of modeling is rigorous mathematically and leads to interesting results. Additional feature of the method is the global affine transformation that helps analyzing distortion and finding a good match by removing a large scale linear disparity between two images.

A New Method for Cartridge Case Image Mosaic

In the process of cartridge case marks detection, due to the limitations of microscope and the unsmoothed specimen surface, not all information can be obtained from just one image. Therefore, two kinds of images can be first obtained and then the information can be supplemented by image mosaic method which facilitates experts’ analysis and the following computer recognition. This paper proposes a new cartridge case image mosaic method by using image registration and fusion techniques. In the registration stage, the initial matching is obtained by using scale invariant feature transform (SIFT), but some incorrect matches greatly affect the registration accuracy. Therefore, in consideration of the specific characteristics of the cartridge case image, graph transformation matching, angle and scale constraint using adaptive K-means clustering are respectively applied to remove incorrect matches. In order to achieve the complementary advantages, voting mechanism is applied to integrate them; meanwhile, genetic algorithm (GA) is employed to select optimal combined parameters, making it possible to adaptively choose to integrate and registration results are optimized based on different images. After refining, the registration accuracy is further enhanced. In the fusion stage, the stitched image is obtained, and histogram matching is employed to smooth visible seams. The mosaic performance is evaluated using visual inspection and objective performance measurements, and results show the advantages of proposed method compared to conventional method.

From engineering diagrams to engineering models: Visual recognition and applications

We present a computational recognition approach to convert network-like, image-based engineering diagrams into engineering models with which computations of interests, such as CAD modeling, simulation, information retrieval and semantic-aware editing, are enabled. The proposed approach is designed to work on diagrams produced using computer-aided drawing tools or hand sketches, and does not rely on temporal information for recognition. Our approach leverages a Convolutional Neural Network (CNN) as a trainable engineering symbol recognizer. The CNN is capable of learning the visual features of the defined symbol categories from a few user-supplied prototypical diagrams and a set of synthetically generated training samples. When deployed, the trained CNN is applied either to the entire input diagram using a multi-scale sliding window or, where applicable, to each isolated pixel cluster obtained through Connected Component Analysis (CCA). Then the connectivity between the detected symbols are analyzed to obtain an attributed graph representing the engineering model conveyed by the diagram. We evaluate the performance of the approach with benchmark datasets and demonstrate its utility in different application scenarios, including the construction and simulation of control system or mechanical vibratory system models from hand-sketched or camera-captured images, content-based image retrieval for resonant circuits and sematic-aware image editing for floor plans.

A Fast Algorithm for the Inverse of a Tridiagonal Period Matrices in Signal Processing

The theory and method of matrix computation, as an important tool, have much important applications such as in computational mathematics, physics, image processing and recognition, missile system design, rotor bearing system, nonlinear kinetics, economics and biology etc. In this paper, Motivated by the references, especially [2], we give the estimates for the lower bounds on the inverse elements of strictly diagonally dominant tridiagonal period matrices.

Estimates for the Lower Bounds on the Inverse Elements of Strictly Diagonally Dominant Tridiagonal Period Matrices in Signal Processing

The theory and method of matrix computation, as an important tool, have much important applications such as in computational mathematics, physics, image processing and recognition, missile system design, rotor bearing system, nonlinear kinetics, economics and biology etc. In this paper, Motivated by the references, especially [2], we give the estimates for the lower bounds on the inverse elements of strictly diagonally dominant tridiagonal period matrices.

Intelligent Intrusion Detection Based on Genetically Tuned Artificial Neural Networks

Artificial intelligence techniques and in particular neural networks (ANN) have been widely employed in an Intrusion Detection System (IDS) design. Due to their high learning ability, their application allows achieving higher performance in various applications. However, they can improve the resource consumption at the same time. This design goal, which is very important in a real life IDS design has received much less attention so far. The paper investigates the optimization methods of improving the ANN-based IDS performance along with the resource consumption. A particular consideration is given to partially connected neural networks that open another way of tuning the ANN structure toward the application. The study examines the choice of the connectivity ratio and its optimization with genetic algorithms. Various genetic algorithms parameters are tested in computer network attacks detection and recognition problems. The results are analyzed and IDS design recommendations are provided.

Identification of vegetable leaf-eating pests based on image analysis

To achieve the computer recognition of vegetable leaf-eating pests and the scientific evaluation of pest level's,the paper proposed a reverse identification method through the eaten leaves to determine the pests.After image pre-processing of the eaten leaves by pests,seven spherical shape feature values of eaten leaves could be automatically extracted: roundness degree,complexity degree,and spherical degree,etc.,and then the BP neural network model for recognition could be built up.The results show that the method is good at recognizing the pest species by the geometric feature of eaten leaves by pests,and give a scientific evaluation for the leaf's harm degree.

A Fast Algorithm for the Inverse of a Class of Tridiagonal Period Matrices in Signal Processing

The Tridiagonal period Matrices, as an important tool, have much important applications ( such as in computational mathematics, physics, image processing and recognition, missile system design, nonlinear kinetics, economics and biology etc). In this paper, we give a fast algorithm for the inverse of the class of tridiagonal period matrices .

Enabling self‐directed computer use for individuals with cerebral palsy: a systematic review of assistive devices and technologies

The purpose of this study was to systematically review published evidence on the development, use, and effectiveness of devices and technologies that enable or enhance self-directed computer access by individuals with cerebral palsy (CP). Nine electronic databases were searched using keywords 'computer', 'software', 'spastic', 'athetoid', and 'cerebral palsy'; the reference lists of articles thus identified were also searched. Thirty articles were selected for review, with 23 reports of development and usability testing of devices and seven evaluations of algorithms to increase computer recognition of input and cursor movements. Twenty-four studies had fewer than 10 participants with CP, with a wide age range of 5 to 77 years. Computer task performance was usually tested, but only three groups sought participant feedback on ease and comfort of use. International standards exist to evaluate effectiveness of non-keyboard devices, but only one group undertook this testing. None of the study designs were higher than American Academy for Cerebral Palsy and Developmental Medicine level IV. Access solutions for individuals with CP are in the early stages of development. Future work should include assessment of end-user comfort, effort, and performance as well as design features. Engaging users and therapists when designing and evaluating technologies to enhance computer access may increase acceptance and improve performance.

Robust speech recognition using improved vector taylor series algorithm for embedded systems

This paper proposes a novel robust speech recognition technique using improved vector Taylor series (VTS) algorithm for embedded systems. It uses a hidden Markov model (HMM) to replace the Gaussian mixture model (GMM) for estimating the clean speech feature, and gives the closed-form solutions of the noise parameters including the mean and variance at each expectation-maximization (EM) iteration. The experimental results show that the proposed algorithm makes a good balance between the computational complexity and recognition accuracy, and thus is more useful for embedded systems.

News on facial views from humans and machine

Abstract Everyday experience suggests that faces can be recognized despite large changes of viewpoint. In this study we used the Inter-Extra-Ortho paradigm from Buelthoff and Edelman (1992) in order to investigate the underlying mechanisms of face recognition. We found systematic effects of viewpoint, which were consistent with computational approaches using interpolation of 2D views. Our results extend the findings from Buelthoff and Edelman on unfamiliar objects to the highly familiar class of faces thus confirming image-based recognition processes independent of class familiarity. In addition, we found that human recognition performance was qualitatively similar to the performance of an extended version of the computational recognition scheme proposed by Wallraven and Buelthoff (2001) using the same faces as in the psychophysical experiments. This algorithm entails processing view-based features and their spatial relations in a dynamic context and is consistent with evidence from psychophysics suggesting separate representations for featural and configural information in face recognition (Schwaninger, Lobmaier, Collishaw, 2002).

Computer recognition systems

The computer recognition systems are nowadays one of the most promising directions in artificial intelligence. This book is the most comprehensive study of this field. It contains a collection of 69 carefully selected articles contributed by experts of pattern recognition. It reports on current research with respect to both methodology and applications. In particular, it includes the following sections: Features, learning and classifiers, Image processing and computer vision, Speech and word recognition, Medical applications, Miscellaneous applications. This book is a great reference tool for scientists who deal with the problems of designing computer pattern recognition systems. Its target readers can be the as well researchers as students of computer science, artificial intelligence or robotics.