Complex Pattern Recognition Research Articles

Magnetic Skyrmion as a Spintronic Deep Learning Spiking Neuron Processor

Deep spiking neural networks (SNNs) have emerged as one of the popular architectures in complex pattern recognition and classification tasks that can be enabled by low-power neuromorphic hardware. However, hardware implementation of such algorithms using conventional CMOS devices is area expensive and energy inefficient. This is owing to the fundamental mismatch between the underlying neuromophic computations and the CMOS transistors along with energy consumption involved in synaptic memory-access operations. Hence, there is a need for novel “neuro-mimetic” devices offering a direct mapping to synaptic and neuronal functionalities together with the possibility of providing in situ synaptic storage. Magnetic skyrmions have recently been proposed as a promising alternative for next-generation information carrier due to remarkably high stability, ultra-low depinning current density, and extremely compact size. In this paper, the design of skyrmion-based devices to emulate biological synapses and neurons is explored, and skyrmionic synapse-based crossbar architectures driving skyrmionic neurons are proposed. We perform a systematic device-circuit-architecture co-design for digit recognition with the MNIST handwritten digits dataset to evaluate the feasibility of our proposal. The device-to-system simulations indicate that the proposed skyrmion-based devices in deep SNNs can potentially achieve two orders of magnitude improvement in energy consumption over an optimized CMOS implementation at a 45 nm technology node.

A Study of Different Classifier Combination Approaches for Handwritten Indic Script Recognition

Script identification is an essential step in document image processing especially when the environment is multi-script/multilingual. Till date researchers have developed several methods for the said problem. For this kind of complex pattern recognition problem, it is always difficult to decide which classifier would be the best choice. Moreover, it is also true that different classifiers offer complementary information about the patterns to be classified. Therefore, combining classifiers, in an intelligent way, can be beneficial compared to using any single classifier. Keeping these facts in mind, in this paper, information provided by one shape based and two texture based features are combined using classifier combination techniques for script recognition (word-level) purpose from the handwritten document images. CMATERdb8.4.1 contains 7200 handwritten word samples belonging to 12 Indic scripts (600 per script) and the database is made freely available at https://code.google.com/p/cmaterdb/. The word samples from the mentioned database are classified based on the confidence scores provided by Multi-Layer Perceptron (MLP) classifier. Major classifier combination techniques including majority voting, Borda count, sum rule, product rule, max rule, Dempster-Shafer (DS) rule of combination and secondary classifiers are evaluated for this pattern recognition problem. Maximum accuracy of 98.45% is achieved with an improvement of 7% over the best performing individual classifier being reported on the validation set.

The significance of tumour microarchitectural features in breast cancer prognosis: a digital image analysis

BackgroundAs only a minor portion of the information present in histological sections is accessible by eye, recognition and quantification of complex patterns and relationships among constituents relies on digital image analysis. In this study, our working hypothesis was that, with the application of digital image analysis technology, visually unquantifiable breast cancer microarchitectural features can be rigorously assessed and tested as prognostic parameters for invasive breast carcinoma of no special type.MethodsDigital image analysis was performed using public domain software (ImageJ) on tissue microarrays from a cohort of 696 patients, and validated with a commercial platform (Visiopharm). Quantified features included elements defining tumour microarchitecture, with emphasis on the extent of tumour-stroma interface. The differential prognostic impact of tumour nest microarchitecture in the four immunohistochemical surrogates for molecular classification was analysed. Prognostic parameters included axillary lymph node status, breast cancer-specific survival, and time to distant metastasis. Associations of each feature with prognostic parameters were assessed using logistic regression and Cox proportional models adjusting for age at diagnosis, grade, and tumour size.ResultsAn arrangement in numerous small nests was associated with axillary lymph node involvement. The association was stronger in luminal tumours (odds ratio (OR) = 1.39, p = 0.003 for a 1-SD increase in nest number, OR = 0.75, p = 0.006 for mean nest area). Nest number was also associated with survival (hazard ratio (HR) = 1.15, p = 0.027), but total nest perimeter was the parameter most significantly associated with survival in luminal tumours (HR = 1.26, p = 0.005). In the relatively small cohort of triple-negative tumours, mean circularity showed association with time to distant metastasis (HR = 1.71, p = 0.027) and survival (HR = 1.8, p = 0.02).ConclusionsWe propose that tumour arrangement in few large nests indicates a decreased metastatic potential. By contrast, organisation in numerous small nests provides the tumour with increased metastatic potential to regional lymph nodes. An outstretched pattern in small nests bestows tumours with a tendency for decreased breast cancer-specific survival. Although further validation studies are required before the argument for routine quantification of microarchitectural features is established, our approach is consistent with the demand for cost-effective methods for triaging breast cancer patients that are more likely to benefit from chemotherapy.

Large-Scale Multimodality Attribute Reduction With Multi-Kernel Fuzzy Rough Sets

In complex pattern recognition tasks, objects are typically characterized by means of multimodality attributes, including categorical, numerical, text, image, audio, and even videos. In these cases, data are usually high dimensional, structurally complex, and granular. Those attributes exhibit some redundancy and irrelevant information. The evaluation, selection, and combination of multimodality attributes pose great challenges to traditional classification algorithms. Multikernel learning handles multimodality attributes by using different kernels to extract information coming from different attributes. However, it cannot consider the aspects fuzziness in fuzzy classification. Fuzzy rough sets emerge as a powerful vehicle to handle fuzzy and uncertain attribute reduction. In this paper, we design a framework of multimodality attribute reduction based on multikernel fuzzy rough sets. First, a combination of kernels based on set theory is defined to extract fuzzy similarity for fuzzy classification with multimodality attributes. Then, a model of multikernel fuzzy rough sets is constructed. Finally, we design an efficient attribute reduction algorithm for large scale multimodality fuzzy classification based on the proposed model. Experimental results demonstrate the effectiveness of the proposed model and the corresponding algorithm.

A NOVEL APPROACH FOR THE PATTERN RECOGNITION OF HAND MOVEMENTS BASED ON EMG AND VPMCD

Essentially, the classification of human hand movements is a process of pattern recognition. However, existing computationally intense and complex pattern recognition methods have failed thus far to be optimally successful in constructing associations between extracted signal features. Due to such limitations, a new pattern recognition method using variable predictive model-based class discrimination (VPMCD) is proposed. This approach considers that the feature values can exhibit inter-relations in nature and such associations will show different forms in different classes. In practice, this is always true for different hand movements. The signals produced by electromyography (EMG) and received from human arm muscles, are characteristically non-linear and non-stationary. A novel hand gesture recognition technique, based on wavelet feature extraction and VPMCD is proposed. First, the maximum values of the wavelet coefficient are extracted as the feature vectors from the surface EMG signals after de-noising. Then, the feature values are regarded as the inputs of the VPMCD classifier. Finally, four movement patterns (hand clenching, hand extension, wrist flexion, and wrist extension) are identified by the outputs of the VPMCD classifier. Our analysis results show that the proposed pattern recognition approach can distinguish different gestures successfully and effectively. Simultaneously, compared with the artificial neural network and the support vector machine classifier, more accurate recognition can be achieved using our proposed technique.

Synaptic Suppression Triplet‐STDP Learning Rule Realized in Second‐Order Memristors

AbstractThe synaptic weight modification depends not only on interval of the pre‐/postspike pairs according to spike‐timing dependent plasticity (classical pair‐STDP), but also on the timing of the preceding spike (triplet‐STDP). Triplet‐STDP reflects the unavoidable interaction of spike pairs in natural spike trains through the short‐term suppression effect of preceding spikes. Second‐order memristors with one state variable possessing short‐term dynamics work in a way similar to the biological system. In this work, the suppression triplet‐STDP learning rule is faithfully demonstrated by experiments and simulations using second‐order memristors. Furthermore, a leaky‐integrate‐and‐fire (LIF) neuron is simulated using a circuit constructed with second‐order memristors. Taking the advantage of the LIF neuron, various neuromimetic dynamic processes, including local graded potential leaking out, postsynaptic impulse generation and backpropagation, and synaptic weight modification according to the suppression triplet‐STDP rule, are realized. The realized weight‐dependent pair‐ and triplet‐STDP rules are clearly in line with findings in biology. The physically realized triplet‐STDP rule is powerful in developing direction and speed selectivity for complex pattern recognition and tracking tasks. These scalable artificial synapses and neurons realized in second‐order memristors can intrinsically capture the neuromimetic dynamic processes; they are the promising building blocks for constructing brain‐inspired computation systems.

Domain Adaptation Using Representation Learning for the Classification of Remote Sensing Images

Traditional machine learning (ML) techniques are often employed to perform complex pattern recognition tasks for remote sensing images, such as land-use classification. In order to obtain acceptable classification results, these techniques require there to be sufficient training data available for every particular image. Obtaining training samples is challenging, particularly for near real-time applications. Therefore, past knowledge must be utilized to overcome the lack of training data in the current regime. This challenge is known as domain adaptation (DA), and one of the common approaches to this problem is based on finding invariant representations for both the training and test data, which are often assumed to come from different “domains.” In this study, we consider two deep learning techniques for learning domain-invariant representations: Denoising autoencoders (DAE) and domain-adversarial neural networks (DANN). While the DAE is a typical two-stage DA technique (unsupervised invariant representation learning followed by supervised classification), DANN is an end-to-end approach where invariant representation learning and classification are considered jointly during training. The proposed techniques are applied to both hyperspectral and multispectral images under different DA scenarios. Results obtained show that the proposed techniques outperform traditional approaches, such as principal component analysis (PCA) and kernel PCA, and can also compete with a fully supervised model in the multispatial scenario.

Multimodal sensor-based weight drop spinal cord impact system for large animals

Background ContextA conventional weight drop spinal cord (SC) impact system for large animals is composed of a high-speed video camera, a vision system, and other things. However, a camera with high speed at over 5,000 frames per second (FPS) is very expensive. In addition, the use of the vision system involves complex pattern recognition algorithms and accurate arrangement of the camera and the target. PurposeThe purpose of this study was to develop a large animal spinal cord injury (SCI) modeling system using a multimodal sensor instead of a high-speed video camera and vision system. Another objective of this study was to demonstrate the possibility of the developed system to measure the impact parameters in the experiments using different stiffness materials and an in vivo porcine SC. Study DesignA multimodal sensor-based SCI impact system was developed for large animals. The experiments to measure SC impact parameters were then performed using three different stiffness materials and a Yucatan miniature pig to verify the performance of system developed. MethodsA comparative experiment was performed using three different stiffness materials such as high-density (HD) sponge, rubber, and clay to demonstrate the system and perform measurement for impact parameters such as impact velocity, impulsive force, and maximally compressed displacement reflecting physical properties of materials. In the animal experiment, a female Yucatan miniature pig of 60-kg weight was used. Impact conditions for all experiments were fixed at freefalling object mass of 50 g and height of 20 cm. ResultsIn the impact test, measured impact velocities were almost the same for the three different stiffness materials at 1.84±0.0153 m/s. Impulsive forces for the three materials of rubber, HD sponge, and clay were 50.88 N, 32.35 N, and 6.68 N, respectively. Maximally compressed displacements for rubber, HD sponge, and clay were 1.93 mm, 3.35 mm, and 15.01 mm, respectively. In the pig experiment, impact velocity, impulsive force, and maximally compressed dural displacement were measured at 1.84 m/s, 13.35 N, and 3.04 mm, respectively. After 3 days from the experiment, paralysis was confirmed for the lower half body of the experimental pig. ConclusionsThrough experiments, it was verified that our proposed system could be used to measure the SC impact parameters and induce SCI for large animals.

Classifier Fusion With Contextual Reliability Evaluation.

Classifier fusion is an efficient strategy to improve the classification performance for the complex pattern recognition problem. In practice, the multiple classifiers to combine can have different reliabilities and the proper reliability evaluation plays an important role in the fusion process for getting the best classification performance. We propose a new method for classifier fusion with contextual reliability evaluation (CF-CRE) based on inner reliability and relative reliability concepts. The inner reliability, represented by a matrix, characterizes the probability of the object belonging to one class when it is classified to another class. The elements of this matrix are estimated from the -nearest neighbors of the object. A cautious discounting rule is developed under belief functions framework to revise the classification result according to the inner reliability. The relative reliability is evaluated based on a new incompatibility measure which allows to reduce the level of conflict between the classifiers by applying the classical evidence discounting rule to each classifier before their combination. The inner reliability and relative reliability capture different aspects of the classification reliability. The discounted classification results are combined with Dempster-Shafer's rule for the final class decision making support. The performance of CF-CRE have been evaluated and compared with those of main classical fusion methods using real data sets. The experimental results show that CF-CRE can produce substantially higher accuracy than other fusion methods in general. Moreover, CF-CRE is robust to the changes of the number of nearest neighbors chosen for estimating the reliability matrix, which is appealing for the applications.

Diversity in Excitation-Inhibition Mismatch Underlies Local Functional Heterogeneity in the Rat Auditory Cortex

Cortical neurons are heterogeneous in their functional properties. This heterogeneity is fundamental for the processing of different features of sensory information. However, functional diversity within a local group of neurons is poorly understood. Here, we demonstrate that neighboring cortical neurons in layer 5 but not those of layer 4 of the rat anterior auditory field (AAF) exhibited a surprisingly high level of diversity in tonal receptive fields. Invivo whole-cell voltage-clamp recordings revealed that the diversity of frequency representation was due to a spectral mismatch between synaptic excitation and inhibition to varying degrees. The spectral distribution of excitation was skewed at different levels, whereas inhibition was homogeneous and non-skewed, similar to the summed spiking activity of local neuronal ensembles, which further enhanced diversity. Our results indicate that AAF in the auditory cortex is involved in processing auditory information in a highly refined manner that is important for complex pattern recognition.

Fault diagnosis of PV array based on optimised BP neural network by improved adaptive genetic algorithm

The fault diagnosis model based on back propagation (BP) neural network is especially suitable for multi-fault and complex pattern recognition. However, the selection of initial weights and thresholds of BP neural network is lack of basis and it is easy to fall into the local optimum. Genetic algorithm (GA) can be used to optimise the initial weights and thresholds of the BP neural network. However, the selection process of GA based on roulette wheel selection is a random operation and the parameters of GA, the crossover probability and the mutation probability, are given a constant value which reduces the efficiency. Thus improved adaptive GA provides improvement in the selection, crossing and mutation process of GA, which prevents the neural network from falling into local optimum more effectively. Results of cases study show that this method can determine the fault types of photovoltaic array effectively.

A spiking neural network for extraction of features in colour opponent visual pathways and FPGA implementation

The human vision system is one of the most complex and the most superior systems that have been known to date. Inspired by the visual ventral pathways, a hierarchical spiking neural network is proposed to extract features from colour opponent visual pathways. The network is composed of multiple visual processing channels to simulate processing mechanism of the visual system. The firing rate map of each channel is recorded and further processed. Useful features are extracted from firing rate map with the characteristics of colour contrast sensitivity, orientation selective and illuminant constancy. These characteristics greatly enrich the expression of colour features and enhance the ability of object recognition. Simulation algorithm of the spiking neural network is based on integrated-and-fire neuron model and a set of receptive fields. Simulation results show that the network has the merit of capability of dealing with complex pattern recognition tasks. Furthermore, a SNN blockset is developed to implement the spiking neural network on FPGAs. It greatly speeds up the hardware modeling of the spiking neural network so that SNNs can be efficiently used in the hardware systems such as robots.

An introduction to learning algorithms and potential applications in geomorphometry and Earth surface dynamics

Abstract. “Learning algorithms” are a class of computational tool designed to infer information from a data set, and then apply that information predictively. They are particularly well suited to complex pattern recognition, or to situations where a mathematical relationship needs to be modelled but where the underlying processes are not well understood, are too expensive to compute, or where signals are over-printed by other effects. If a representative set of examples of the relationship can be constructed, a learning algorithm can assimilate its behaviour, and may then serve as an efficient, approximate computational implementation thereof. A wide range of applications in geomorphometry and Earth surface dynamics may be envisaged, ranging from classification of landforms through to prediction of erosion characteristics given input forces. Here, we provide a practical overview of the various approaches that lie within this general framework, review existing uses in geomorphology and related applications, and discuss some of the factors that determine whether a learning algorithm approach is suited to any given problem.

Feature selection via binary simultaneous perturbation stochastic approximation

Feature selection (FS) has become an indispensable task in dealing with today’s highly complex pattern recognition problems with massive number of features. In this study, we propose a new wrapper approach for FS based on binary simultaneous perturbation stochastic approximation (BSPSA). This pseudo-gradient descent stochastic algorithm starts with an initial feature vector and moves toward the optimal feature vector via successive iterations. In each iteration, the current feature vector’s individual components are perturbed simultaneously by random offsets from a qualified probability distribution. We present computational experiments on datasets with numbers of features ranging from a few dozens to thousands using three widely-used classifiers as wrappers: nearest neighbor, decision tree, and linear support vector machine. We compare our methodology against the full set of features as well as a binary genetic algorithm and sequential FS methods using cross-validated classification error rate and AUC as the performance criteria. Our results indicate that features selected by BSPSA compare favorably to alternative methods in general and BSPSA can yield superior feature sets for datasets with tens of thousands of features by examining an extremely small fraction of the solution space. We are not aware of any other wrapper FS methods that are computationally feasible with good convergence properties for such large datasets.

Automatic Colorectal Polyp Detection in Colonoscopy Video Frames

Colonoscopy is currently the best technique available for the detection of colon cancer or colorectal polyps or other precursor lesions. Computer aided detection (CAD) is based on very complex pattern recognition. Local binary patterns (LBPs) are strong illumination invariant texture primitives. Histograms of binary patterns computed across regions are used to describe textures. Every pixel is contrasted relative to gray levels of neighbourhood pixels. In this study, colorectal polyp detection was performed with colonoscopy video frames, with classification via J48 and Fuzzy. Features such as color, discrete cosine transform (DCT) and LBP were used in confirming the superiority of the proposed method in colorectal polyp detection. The performance was better than with other current methods.

Implementing Multilayer Neural Network Behavior Using Polychronous Wavefront Computation

Polychronous Wavefront Computation (PWC) is an abstraction of spiking neural networks that provides a potentially practical model for implementing neuromorphic computing systems. While it's has been shown to exhibit some basic computational capabilities, its use in complex neuro-computational models remains to be explored. The paper presents a model and approach for configuring PWC transponders to implement multilayer neural network behavior to provide a basis for more complex applications of the technology. The model uses a set of input transponders representing pattern features to stimulate hidden layer transponders that combine features and trigger output layer transponders to identify patterns. The input layer transponder geometry is selected to create wavefront intersections for all relevant feature combinations. Hidden layer transponders are positioned by solving the intersection of the circles equations defined by sets of input transponders. Output layer transponders are defined to collect complete sets of features for recognition based on the hidden layer transponder geometry. The approach uses the intersections of three wavefronts to maximize transponder selectivity and increase information density. The concept is experimentally demonstrated and analyzed with a 7-segment display digit recognition application which provides a simple but representative example of more complex pattern recognition problems.

Toward Big Data Analytics

Diabetes is one of the top priorities in medical science and health care management, and an abundance of data and information is available on these patients. Whether data stem from statistical models or complex pattern recognition models, they may be fused into predictive models that combine patient information and prognostic outcome results. Such knowledge could be used in clinical decision support, disease surveillance, and public health management to improve patient care. Our aim was to review the literature and give an introduction to predictive models in screening for and the management of prevalent short- and long-term complications in diabetes. Predictive models have been developed for management of diabetes and its complications, and the number of publications on such models has been growing over the past decade. Often multiple logistic or a similar linear regression is used for prediction model development, possibly owing to its transparent functionality. Ultimately, for prediction models to prove useful, they must demonstrate impact, namely, their use must generate better patient outcomes. Although extensive effort has been put in to building these predictive models, there is a remarkable scarcity of impact studies.

HfO<sub>2</sub>-Based OxRAM Devices as Synapses for Convolutional Neural Networks

In this paper, the use of HfO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -based oxide-based resistive memory (OxRAM) devices operated in binary mode to implement synapses in a convolutional neural network (CNN) is studied. We employed an artificial synapse composed of multiple OxRAM cells connected in parallel, thereby providing synaptic efficacies. Electrical characterization results show that the proposed HfO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -based OxRAM technology offers good electrical properties in terms of endurance (>10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">8</sup> cycles), speed (<;10 ns), and low energy (<;10 pJ), and thus being well suited for neuromorphic applications. A device physical model is developed in order to study the variability of the resistance as a function of the stochastic position of oxygen vacancies in 3-D. Finally, the proposed binary OxRAM synapse has been used for CNN system-level simulations. High accuracy (recognition rate > 98%) is demonstrated for a complex visual pattern recognition application. We demonstrated that the resistance variability and the reduced memory window of the OxRAM cells when operated at extremely low programming conditions (<;10 pJ per switching event) have a small impact on the performances of proposed OxRAM-based CNN (recognition rate 94%).

Emerging Trends in Design and Applications of Memory-Based Computing and Content-Addressable Memories

Content-addressable memory (CAM) and associative memory (AM) are types of storage structures that allow searching by content as opposed to searching by address. Such memory structures are used in diverse applications ranging from branch prediction in a processor to complex pattern recognition. In this paper, we review the emerging challenges and opportunities in implementing different varieties of CAM/AM structures. Beyond-CMOS silicon and nonsilicon memory technologies hold significant promise in implementing dense, fast, and energy-efficient CAM/AM structures. We describe circuit/architecture level implementations of CAM/AM using these technologies, as well as novel applications in different domains, including informatics, text analytics, data mining, and reconfigurable computing platforms.

Oxide-Based Synaptic Transistors for Neuromorphic Systems

Inspired by the energy-efficient and cognitive computational ability of a biological brain, neuromorphic computing is an attractive computation paradigm that complements the von Neumann architecture. Hardware implementations of synapses, dendrites and spiking neural networks represent a promising computational paradigm for solving complex pattern recognition and sensory processing tasks. Here, artificial electronic synapses and dendrites are proposed based on the oxide-based electrical-double-layer transistors. Strong electrostatic coupling effect and electrochemical doping were observed due to proton migration in the P-doped SiO2 electrolyte films. Protons in the P-doped SiO2 electrolyte and oxide channel conductance were regarded as the neurotransmitter and synaptic weight, respectively. Spike-timing dependent plasticity, short-term plasticity, including paired-pulse facilitation, dynamic filtering was mimicked. Most importantly, dendritic arithmetic and spiking pH sensors with extremely low energy dissipation of 0.5pJ are also demonstrated in a simple artificial synapse with multiple pre-synaptic inputs. Our oxide-based protonic/electronic hybrid artificial synaptic transistors are potential building blocks for brain-inspired computers and neuromorphic systems. [1] L. Q. Zhu, C. J. Wan, L. Guo, Y. Shi, and Q. Wan*. Nature Communications, 5, 3158, 2014. [2] C. J. Wan, L. Q. Zhu, J. M. Zhou, Y. Shi, Q. Wan*. Nanoscale. 6, 4491, 2014. [3] C.J. Wan, L. Q. Zhu, Y. H. Liu, Y. Shi, and Q. Wan*. IEEE Electron Device Letters. 35, 672, 2014. [4] L.Q. Guo, Q. Wan*, C. J. Wan, L. Q. Zhu, Y. Shi. IEEE Electron Device Letters. 34,1581, 2013. [5] M.Z. Dai, Q. Wan*. Nano Letters 11, 3987, 2011.