Pattern recognition system: from classical methods to deep learning techniques

Purpose:This study aims to compare the accuracy of the ADNEX MR scoring system and pattern recognition system to evaluate adnexal lesions indeterminate on the US exam. Methods:In this cross-sectional retrospective study, pelvic DCE-MRI of 245 patients with 340 adnexal masses was studied based on the ADNEX MR scoring system and pattern recognition system. Results:ADNEX MR scoring system with a sensitivity of 96.6% and specificity of 91% has an accuracy of 92.9%. The pattern recognition system’s sensitivity, specificity, and accuracy are 95.8%, 93.3%, and 94.7%, respectively. PPV and NPV for the ADNEX MR scoring system were 85.1 and 98.1, respectively. PPV and NPV for the pattern recognition system were 89.7% and 97.7%, respectively. The area under the ROC curve for the ADNEX MR scoring system and pattern recognition system is 0.938 (95% CI, 0.909-0.967) and 0.950 (95% CI, 0.922-0.977). Pairwise comparison of these AUCs showed no significant difference (p = 0.052). Conclusion:The pattern recognition system is less sensitive than the ADNEX MR scoring system, yet more specific.

The authors demonstrate how a pattern-recognition system can be applied to the interpretation of capacitance-voltage (C-V) curves on an MOS test structure. By intelligently sequencing additional measurements it is possible to accurately extract the maximum amount of information available from C-V and conductance-voltage (G-V) measurements. The expert system described, (CV-EXPERT), is completely integrated with the measurement, instrumentation, and control software and is thus able to call up a sequence of individually tailored tests for the MOS test structure under investigation. The prototype system is able to correctly identify a number of process faults, including a leaky oxide, as shown. Improvements that could be gained from developing rules to coordinate G-V, capacitance-time, and doping profile measurements simply by recognizing the important factors in the initial C-V measurement are illustrated. >

Optimal Speech feature extraction which tries to acquire a parametric illustration of an input speech signal plays a vital role in the overall performance of an Automatic Speech Recognition (ASR) system. A good feature extraction technique along with feature selection algorithm should capture the important features of the signal and also should reject some irrelevant features. Feature selection is a critical task which can influence the performance of pattern classification and recognition system. In this paper, we have presented a hybrid evolutionary algorithm called Artificial Bee Colony (ABC) and Particle Swarm Optimization (PSO) for optimal feature extraction and selection strategy. The proposed method consists of three stages, preprocessing, feature extraction and selection, and recognition. Initially, the preprocessing is done by wiener filter to reduce the noise level. Next, we extract eight type of statistical and acoustic features in the feature stage. The optimal set of extracted features is selected by using hybrid ABC-PSO algorithm. Finally, these optimized features are used for training the Support Vector Machine (SVM) classifier and based on these optimized features of the given input speech signal the corresponding text is displayed as the output. The proposed ASR is implemented on the MATLAB working platform and the experimental results show that overall performance of the system is high and the proposed hybrid algorithm is best suited for speech recognition.

Reducing the effect of wrist variation on pattern recognition of Myoelectric Hand Prostheses Control through Dynamic Time Warping

This paper proposes a prototype model of robust iris pattern recognition (PR) system for classification of ten different persons using neural network. Feature extraction algorithms are developed and an optimal feature vector comprising of features in relation to image statistics, texture and 2-D transform domain is formed. It is observed that 2D Walsh Hadamard Transform (WHT) entails the best performance as compared to other image transforms. Different neural network configurations, such as, Multi Layer Perceptron (MLP), Radial Basis Function (RBF) and Support Vector Machine (SVM) are implemented after systematically varying the concerned parameters of the respective networks and MLP with single hidden layer is seen to outperform all others with respect to performance on cross validation dataset derived from CASIA iris image database. Further, the sensitivity analysis is carried out over this network in order to reduce time and space complexity of the network. To make this network more robust; controlled Gaussian and Uniform noise are injected in all input features and it is noticed that for both types of noise, the proposed PR system can sustain the variance up to 0.1.

The technique of identifying patterns with the aid of a machine learning system is called pattern recognition. The classification of data based on previously acquired knowledge or on statistical data extrapolated from patterns and/or their representation is known as pattern recognition. Pattern recognition is such pattern where we scale some object and it is a technique critical in many areas, including surveillance cameras, access control systems, biometric data, interactive game apps, human computer interaction. Through this article, we explain the application of a multi-pattern recognition framework in various steps and used the classification framework to identify the object intensity using machine learning. Our study is also based on some parameters where we examine the results between recognition system and ML technique. We conclude a proposed system for the implementation of pattern recognition system and this work is also useful for 3D image preprocessing as well as artificial neural networks to improve the system's recognition rate.

Feature extraction is an important factor to improve the recognition rate of coal and gangue. The existing feature extraction methods have some shortcomings in coal and gangue recognition, such as unsatisfactory recognition rate for actual images. Therefore, aiming at the shortcomings of existing coal and gangue feature extraction methods in coal and gangue recognition, the contour feature extraction method is researched based on images analysis to coal and gangue recognition. The center normalization detrended fluctuation analysis (CNDFA) feature extraction algorithm of contour is proposed based on the process of contour feature extraction for coal and gangue. Based on the analysis of the representative features of coal and gangue, the extraction process of target contour features is established based on hardness difference. Combined with the overall features of contour curve after detrending, a center normalized CNDFA feature extraction algorithm is proposed. First, the detrended analysis of contour curve is realized by least square optimal fitting, and then the detrended data are normalized. Finally, the contour features are described quantitatively by multifractal (MF) spectrum to form the geometric features of the target contour curve, which is used to train the support vector machine classifier. The experiment is carried out on the basis of image preprocessing, and the CNDFA method and other feature extraction methods, such as wavelet, gray level co-occurrence matrix, gray level difference statistics, auto-correlation function, and MF, are applied to the contour feature extraction of coal and gangue. Through the comprehensive comparison of the results after different methods recognition in confusion matrix, accuracy, and coal cleanliness, it is concluded that the overall effect of the CNDFA method is better than other methods, and the accuracy is improved by 5% to 25%. The results show that the CNDFA method has better performance. Compared with the other methods, it can better extract the contour features to improve the recognition rate of coal and gangue.

A pattern recognition system is designed to solve the model characterization problem of distributed systems. The characterization problem in the pattern recognition domain is the mapping of the observation and input data of the distributed system into one of the designated classes. Mathematical structures which are likely to represent the system are designated as classes from the a priori information of the system under consideration. A feature extractor and a classifier are designed which are tailored for this problem. The pattern recognition system is assessed by using the leaving-one-out method on simulated data.

A novel approach for water quality classification based on the integration of deep learning and feature extraction techniques

We introduce a system for machine recognition of music patterns. The problem is put into a pattern recognition framework in the sense that an error between a target pattern and scanned pattern is minimized. The error takes into account pitch and rhythm information. The pitch error measure consists of an absolute (objective) error and a perceptual error. The latter depends on an algorithm for establishing the tonal context which is based on Krumhansl's (1990) key-finding algorithm. The sequence of maximum correlations that it outputs is smoothed with a cubic spline and is used to determine weights for perceptual and absolute pitch errors. Maximum correlations are used to create the assigned key sequence, which is then filtered by a recursive median filter to improve the structure of the output of the key finding algorithm. A procedure for choosing weights given to pitch and rhythm errors is discussed.

Classical feature extraction and data projection methods have been well studied in the pattern recognition and exploratory data analysis literature. We propose a number of networks and learning algorithms which provide new or alternative tools for feature extraction and data projection. These networks include a network (SAMANN) for J.W. Sammon's (1969) nonlinear projection, a linear discriminant analysis (LDA) network, a nonlinear discriminant analysis (NDA) network, and a network for nonlinear projection (NP-SOM) based on Kohonen's self-organizing map. A common attribute of these networks is that they all employ adaptive learning algorithms which makes them suitable in some environments where the distribution of patterns in feature space changes with respect to time. The availability of these networks also facilitates hardware implementation of well-known classical feature extraction and projection approaches. Moreover, the SAMANN network offers the generalization ability of projecting new data, which is not present in the original Sammon's projection algorithm; the NDA method and NP-SOM network provide new powerful approaches for visualizing high dimensional data. We evaluate five representative neural networks for feature extraction and data projection based on a visual judgement of the two-dimensional projection maps and three quantitative criteria on eight data sets with various properties.

Bio-Inspired Hybrid Intelligent Systems for Image Analysis and Pattern Recognition comprises papers on diverse aspects of bio-inspired models, soft computing and hybrid intelligent systems. The articles are divided into four main parts. The first one consists of papers that propose new fuzzy and bio-inspired models to solve general problems. The second part deals with the main theme of modular neural networks in pattern recognition, which are basically papers using bio-inspired techniques. The third part contains papers that apply hybrid intelligent systems to the problem of time series analysis and prediction, while the fourth one shows papers dealing with bio-inspired models in optimization and robotics applications. An edited book in which both theoretical and application aspects are covered.

A hybrid evolutionary learning algorithm is presented that synthesizes a complete multiclass pattern recognition system. The approach uses a multifaceted representation that evolves layers of processing to perform feature extraction from raw input data, select cooperative sets of feature detectors, and assemble a linear classifier that uses the detectors' responses to label targets. The hybrid algorithm, called hybrid evolutionary learning for pattern recognition (HELPR), blends elements of evolutionary programming, genetic programming, and genetic algorithms to perform a search for an effective set of feature detectors. Individual detectors are represented as expressions composed of morphological and arithmetic operations. Starting with a few small random expressions, HELPR expands the number and complexity of the features to produce a recognition system that achieves high accuracy. Results are presented that demonstrate the performance of HELPR-generated recognition systems applied to the task of classification of high-range resolution radar signals.

The oil and gas industry need for high performing and low cost multiphase meters is ever more justified given the rapid depletion of conventional oil reserves that has led oil companies to develop smaller and marginal fields and reservoirs in remote locations and deep offshore, thereby placing great demands for compact and more cost effective solutions of on-line continuous multiphase flow measurement for well testing, production monitoring, production optimisation, process control and automation. The pattern recognition approach for clamp-on multiphase measurement employed in this study provides one means for meeting this need. High speed caesium-137 radioisotope-based densitometers were installed vertically at the top of a 50.8mm and 101.6mm riser as well as horizontally at the riser base in the Cranfield University multiphase flow test facility. A comprehensive experimental campaign comprising flow conditions typical of operating conditions found in the Petroleum Industry was conducted. The application of a single gamma densitometer unit, in conjunction with pattern recognition techniques to determine both the phase volume fractions and velocities to yield the individual phase flow rates of horizontal and vertically upward multiphase flows was investigated. The pattern recognition systems were trained to map the temporal fluctuations in the multiphase mixture density with the individual phase flow rates using statistical features extracted from the gamma counts signals as their inputs. Initial results yielded individual phase flow rate predictions to within ±5% relative error for the two phase airwater flows and ±10% for three phase air-oil-water flows data.

With the development of industrial automation, there is a growth in number of vehicles resulted to demand of parking space, which is costly in metropolitan areas. For finding vacant parking facilities, drivers spend more time on roads which causes additional fuel consumption, traffic congestion, and pollution. For addressing parking issues, this study presents a novel framework based on real-time monitoring and image recognition technique. The major contribution of this study is threefold: Initially, analyze the input images which are composed from the event recorders in currently driving cars toward determining the parking spaces availability. Second, this image is preprocessed using a hybrid algorithm, which is a combination of feature extraction (local binary pattern) and pattern recognition (Bayes classifier) technique. With the help of image recognition technique, it can track the availability of parking spaces with images on weather condition. Also, image filtering technique is applied to remove noisy information; hence, we can detect the park lots in any weather condition. Third, consider the utilization of parking facilities, distance to the recommended parking facility, and time to reach from source to destination. Finally, the performance of the suggested technique is validated by the measure of classification accuracy, precision, recall, and f-measure.KeywordsVehicle detectionImage processingSmart parking systemFeature extraction and classification