New Publications are available for Pattern recognition
http://dl-live.theiet.org
New Publications are available now online for this publication.
Please follow the links to view the publication.Development of an automated system for chromosome classification
http://dl-live.theiet.org/content/conferences/10.1049/ic.2009.0153
Automated chromosome classification has been an important pattern recognition problem for decades. In this paper a novel method for Karyotyping (segmentation and classification) of chromosomes is used. Segmentation is carried out using watershed algorithm and area of each of the chromosomes is calculated to classify them in different classes. Though multiplex fluorescent in situ hybridization (MFISH) is recently very commonly used technique for classification, it is very costly. The method proposed in this paper is applicable to gray and color images as well. The algorithm was tested on samples from genetic labs and has shown acceptable classification results. (8 pages)Pit pattern classification of zoom-endoscopic colon images using wavelet texture features
http://dl-live.theiet.org/content/conferences/10.1049/cp_20060361
Wavelet-based techniques for an automated classification of magnifying endoscope images with respect to pit patterns of colon lesions are discussed and compared. Algorithms classically used for wavelet-based classification but developed originally for texture classification applications provide encouraging results but still lack of the accuracy required for sensible clinical usage. (4 pages)Multiple cell-particle tracking based on multi-resolution block-matching using genetic algorithm
http://dl-live.theiet.org/content/conferences/10.1049/cp_20060375
The regulation of cargo trafficking between intracellular compartments is central to the control of many cellular processes. Rapid sub-second imaging shows that GFP-GLUT4 vesicles exhibited two types of movement, firstly rapid vibrations around a point and linear movements. To gain further insights into the molecular basis underlying these movements, and understand how insulin causes translocation to the cell surface, it is essential to develop methods by which we can accurately track and measure the speed and directionality of vesicles, and their fluorescent intensity, under a variety of conditions. In this paper we propose a novel tracking mechanism to study the dynamics of these vesicles in single living cells. The proposed model is based on a multi-resolution block matching algorithm for motion estimation and a genetic algorithm. We demonstrate the success of the model through results obtained on various real time videos. (4 pages)Improving prostate cancer diagnosis using tabu search
http://dl-live.theiet.org/content/conferences/10.1049/cp_20040578
The introduction of multispectral imaging in pathology problems such as the identification of prostatic cancer is recent. Unlike conventional RGB color space, it allows the acquisition of a large number of spectral bands within the visible spectrum. This results in a feature vector of size greater than 100. For such high dimensionality problems, pattern recognition techniques suffer from the well-known curse-of-dimensionality problem. The two well known techniques to solve this problem are feature extraction and feature selection. A feature selection technique using tabu search with an intermediate-term memory is proposed. The cost of a feature subset is measured by the leave-one-out correct-classification rate of a nearest-neighbor (1-NN) classifier. Experiments have been carried out on prostate cancer textured multispectral images and the results have been compared with a reported classical feature extraction technique. Results have indicated a significant boost in the performance in terms of both minimizing features and maximizing classification accuracy.Contrast enhancement to improve tracking and classification performance
http://dl-live.theiet.org/content/conferences/10.1049/ic_19980424
This paper describes the effects of using a contrast enhancement technique in image processing applications relating to target identification, tracking and classification. Specifically, it records the evaluation processes and the improvement provided by the technique in these applications when using imagery taken from the infrared waveband. Previous work has shown that a significant improvement in effective visibility may be achieved by correcting for the atmospheric degradation apparent in images taken from the visible and infrared wavebands. The enhancement technique is based on a three parameter model for the transmission of flux from the terrain to the imaging sensor. The measured flux is given by: I=C<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">0</sub>(1+C<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">1</sub>exp(-σR)) where R is the range and C<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">0</sub>, C<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">1</sub> and σ are constants which depend on the atmospheric parameters and the terrain radiance, and σ is known as the extinction coefficient. Using actual image data, estimates for the three parameters can be calculated. When the model is combined with a range map corresponding to the image, the effects of atmospheric degradation may be reversed. The result is an improvement in the range over which targets can be detected together with increased object clarity. Any image processing method which increases the detectable range and clarity of objects improves the capability and performance of applications concerning target classification and tracking. (3 pages)Application of neural network pattern recognition to the geometrical characterisation of optical fibres
http://dl-live.theiet.org/content/conferences/10.1049/cp_19950741
Automation of the image shearing measurement technique applied to optical fibre geometry is discussed. The use of both edge detection filters and a multi-layer perceptron neural network to identify the critical `just touch' condition are compared. The repeatability of cladding diameter measurements using both methods are presented which demonstrate the superiority of the neural network technique.Human gait recognition in canonical space using temporal templates
http://dl-live.theiet.org/content/journals/10.1049/ip-vis_19990187
A system for automatic gait recognition without segmentation of particular body parts is described. Eigenspace transformation (EST) has already proved useful for several tasks including face recognition, gait analysis, etc; it is optimal in dimensionality reduction by maximising the total scatter of all classes but is not optimal for class separability. A statistical approach that combines EST with canonical space transformation (CST) is proposed for gait recognition using temporal templates from a gait sequence as features. This method can be used to reduce data dimensionality and to optimise the class separability of different gait sequences simultaneously. Incorporating temporal information from optical-flow changes between two consecutive spatial templates, each temporal template extracted from computation of optical flow is projected from a high-dimensional image space to a single point in a low-dimensional canonical space. Using template matching, recognition of human gait becomes much faster and simpler in this new space. As such, the combination of EST and CST is shown to be of considerable potential in an emerging new biometric.Recognition of various tactile stimuli using independent component analysis and <i xmlns="http://pub2web.metastore.ingenta.com/ns/">k</i>-means
http://dl-live.theiet.org/content/journals/10.1049/iet-spr.2009.0131
A self-developed integrated system is employed to record and analyse intracortical evoked potentials from the primary somatosensory cortex of rats. Four different neural signals are recorded under no stimulation and stimulation using a toothbrush, pen shaft and toothpick separately. These evoked signals undergo preprocessing and post-processing, in that order. In order to improve the shortcoming of independent component analysis (ICA), which the magnitude and sequence of estimated independent components are ambiguous. The authors propose the dynamic dimension increasing method to form a feature vector by correlation coefficient matrix and mitigate the drawback of ICA. Then, <i xmlns="http://pub2web.metastore.ingenta.com/ns/">k</i>-means is employed to group the feature vector into different clusters. The authors use the information of monitoring subsystem to check the experimental results by using a video recording device. Finally, the presented methods are utilised to extract the features from various evoked potentials and distinguish the stimulants from different sensory signals.Face recognition from synchronised visible and near-infrared images
http://dl-live.theiet.org/content/journals/10.1049/iet-spr.2008.0173
The improvement in face recognition that can be obtained from the simultaneous availability of normal and near-infrared (NIR) images is quantitatively measured. The authors designed a camera that can perform a simultaneous acquisition of a NIR image and visible images (VI) and therefore built a face database with this camera aiming at comparing the performance of several algorithms on both types of images when illumination variations occur. The authors noticed the stability of the performance of all the tested algorithms on infrared images upon illumination variation, and the improvement in performance that results from the fusion of these two different types of images.Robust detection of horizontal small targets using synergistic spatial filtering
http://dl-live.theiet.org/content/journals/10.1049/el.2009.0795
A robust noise reduction and background estimation method is presented to detect small targets around the horizon for infrared search and track. Directional background estimation and removal after a double window filter is very synergistic in terms of detection rate and false alarms. Experimental results validate the robustness of the proposed method.Two new beta-related probability density functions
http://dl-live.theiet.org/content/journals/10.1049/el_19981593
Two new probability density functions (PDFs) closely related to the beta PDF are derived. The first PDF generalises the well-known relation about gamma and beta random variables (RVs). The second enables the modelling of uncertainty in one of the parameters of the beta RV. Both functions have a number of interesting applications in image processing, radar and pattern recognition.Texture analysis of fluorescence microscopic images of colonic tissue sections
http://dl-live.theiet.org/content/journals/10.1049/mbec_20013562
The aim of this study was to assess the potential of texture analysis for the characterization of fluorescence images from colonic tissue sections stained with a novel and selective fluoroprobe, Rhodamine B-phenylboronic acid. Fluorescence microscopy images of colonic healthy mucosa (n=35) and adenocarcinomas (n=35) were digitally captured and subjected to image texture analysis. Textural features derived from the grey level co-occurrence matrix were calculated. A modified version of the multiple discriminant analysis criterion was used to choose an appropriate subset of features. A minimum Mahalanobis distance, linear discriminant classifier and a simple evaluation ‘score’ method were used to classify image feature data into the two categories. A subset of four textural features was selected and used for the description and classification of each image field. They were found appropriate to correctly classify 95% of the images into the two classes, using two different classifiers. These features contained information about local homogeneity and grey level linear dependencies of the image. This study demonstrated that texture analysis techniques could provide valuable diagnostic decision support in a complex domain such as colorectal tissue.Deformable contour method based on variational approach to a constrained optimisation formulation
http://dl-live.theiet.org/content/journals/10.1049/el_20040081
A deformable contour method derived from the variational approach to a constrained contour energy minimisation formulation is presented. The constraint is a function that characterises target object interiors. A new constraint is also proposed with better results when compared to other conventional deformable contour methods.Optimisation procedures for diagnostic processing of hand-drawn geometric figures
http://dl-live.theiet.org/content/journals/10.1049/el_20030121
An investigation into task and feature selection for analysing diagnostic drawings using handwriting dynamics and image processing techniques is presented. Performance across a standard test battery and a subset selection process for both features and tasks is examined. It is shown that a reduced task domain provides results which can identify trends in patient performance at a higher accuracy than is obtainable when the whole feature/task set is used.Segmentation method for face detection in complex background
http://dl-live.theiet.org/content/journals/10.1049/el_20000233
A new method for segmenting candidate regions of human faces with a complex background is proposed. A number of evolutionary agents are distributed in the 2-D colour image environment in order to find the face-like pixels and locate each face-like region. Experimental results show that the proposed method is fast and robust.Application of analogue self-electro-optic effect device with five input photodetectors to edge detection
http://dl-live.theiet.org/content/journals/10.1049/el_20001146
A second order two-dimensional spatial image differentiation has been performed using an analogue self-electro-optic effect device (SEED). The device integrates a pair of quantum-well modulators with a set of five conventional photodiodes designed to extract local spatial differentiation. The result shows that this device is suitable for studying early vision processes.Minimum stable convergence criteria for Stochastic Diffusion Search
http://dl-live.theiet.org/content/journals/10.1049/el_20040096
An analysis of Stochastic Diffusion Search (SDS), a novel and efficient optimisation and search algorithm, is presented, resulting in a derivation of the minimum acceptable match resulting in a stable convergence within a noisy search space. The applicability of SDS can therefore be assessed for a given problem.Inter-chromosome texture as a feature for automatic identification of metaphase spreads
http://dl-live.theiet.org/content/journals/10.1049/mbec_20023684
This paper reports results for a new measure of texture coarseness, as a step towards automation of metaphase finding in cell-proliferation studies. This measure is highly specific to grey-level inter-chromosome coarseness features in microscopic images of metaphase spreads and allows the texture quantification of cytological objects, analysing the intensity profile between chromosome-extrema samples. Chromosome fragments produce patterns of pixels at low resolution, and the local neighbourhood of their individual extrema presents a characteristic coarseness along intensity profiles, on randomly oriented test lines. Results of the use of this new measure on microscope images of fields of metaphases and artifacts are compared with some representative texture measures and the performance of reported metaphase finders. This new measure outperforms the latter, when applied in metaphase detection and elimination of artifacts. This coarseness feature provides a specific metaphase signature that can be used in conjunction with other morphological and textural parameters for automated metaphase discrimination.Local intensity distribution descriptor for object detection
http://dl-live.theiet.org/content/journals/10.1049/el.2010.3256
A novel local intensity distribution (LID) descriptor for object detection is proposed. By capturing the distribution of local intensity differences effectively, the LID descriptor is insensitive to illumination changes while being more compact and discriminative compared with the popular local binary pattern. Two LID descriptors can be efficiently compared using the Kullback-Leibler distance. The efficacy and efficiency of the proposed descriptor have been verified in the task of detecting humans from static images.Selection of image classifiers
http://dl-live.theiet.org/content/journals/10.1049/el_20000374
An approach to classifier combination based on the concept of ‘dynamic classifier selection’ is presented. The results reported show that the proposed approach enables effective image classification systems to be developed.Parameterised structured light imaging for depth edge detection
http://dl-live.theiet.org/content/journals/10.1049/el.2010.2791
This reported research features parameterised structured light imaging that is practically useful for detecting depth edges. Given input parameters such as the range of distances of an object from the camera/projector and minimum detectable depth difference, the presented method is capable of computing an optimal pattern width and the number of structured light images that are needed to detect all depth edges in the specified range of distances that have at least the given detectable depth difference. Application of this parameter control to the detection of silhouette edges for visual hull reconstruction shows the effectiveness of the method.4.65 Gbit/s optical four-bit pattern matching using silica-based waveguide circuit
http://dl-live.theiet.org/content/journals/10.1049/el_19901134
An ultra-fast coherent optical pattern matching circuit fabricated with silica-based waveguides on a silicon substrate is presented. This circuit can detect any specified pattern in an ultra-fast optical signal sequence. Experimental four bit pattern matching is successfully demonstrated at a bit-rate of 4.65 Gbit/s.Face recognition: combining cognitive psychology and image engineering
http://dl-live.theiet.org/content/journals/10.1049/ecej_19920050
Many cognitive tasks that are easy for humans to perform are proving difficult to emulate in computer systems. Combining the disciplines of psychology and engineering may offer a solution to some of these problems. A connectionist or neural network model of face recognition by humans which incorporates aspects of a model proposed by cognitive psychologists is presented. A comparative set of experiments has been performed using this simulation and human subjects for familiar face recognition. By employing the same stimuli for both humans and the computer model, it is possible to advance not only our understanding of human cognition but also to develop improved automated systems for face recognition.Phase-dominant spatial light modulators
http://dl-live.theiet.org/content/journals/10.1049/el_19880424
Computer simulation results are given for an optical correlator in which the spatial light modulator has a strong phase modulation and a weak proportional amplitude modulation at each pixel. The signal/noise ratio on the correlation plane is found to be much higher than that of a matched filter, and the residual signal-dependent amplitude modulation improves the system performance in some cases.Optical correlation in Bi<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">12</sub>SiO<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">20</sub> at 632.8 nm
http://dl-live.theiet.org/content/journals/10.1049/el_19840023
Optical image correlation with a low-power HeNe laser has been demonstrated using the phenomenon of degenerate four-wave mixing in Bi<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">12</sub>SiO<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">20</sub>. Preliminary results for character recognition, using a compact two-dimensional image processor, are presented.Erratum: Optical correlation in Bi<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">12</sub>SiO<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">20</sub> at 632.8 nm
http://dl-live.theiet.org/content/journals/10.1049/el_19840181
Computer simulation of hybrid cross-correlators
http://dl-live.theiet.org/content/journals/10.1049/ip-j.1986.0005
A computer model for simulating hybrid optical/digital cross correlators for pattern recognition is developed to investigate accuracy and limitations of a hybrid cross correlator compared to an ideal digital cross-correlator. The hybrid system is based on matrix-addressed <strong xmlns="http://pub2web.metastore.ingenta.com/ns/">LCD</strong> displays and <strong xmlns="http://pub2web.metastore.ingenta.com/ns/">CCD</strong> camera chips as interface devices. Both joint-transform and frequency-plane correlators can be simulated, and different physical distortion effects from the optical components and from the interface devices can be included. The evaluation of the hybrid system is carried out in the correlation plane, which is the output of the system.Partial discharge pattern classification using multilayer neural networks
http://dl-live.theiet.org/content/journals/10.1049/ip-a-3.1993.0049
Partial discharge measurement is an important means of assessing the condition and integrity of insulation systems in high voltage power apparatus. Commercially available partial discharge detectors display them as patterns by an elliptic time base. Over the years, experts have been interpreting and recognising the nature and cause of partial discharges by studying these patterns. A way to automate this process is reported by using the partial discharge patterns as input to a multilayer neural network with two hidden layers. The patterns are complex and can be further complicated by interference. Therefore the recognition process appropriately qualifies as a challenging neural network task. The simulation results, and those obtained when tested with actual patterns, indicate the suitability of neural nets for real world applications in this emerging domain. Some limitations of this method are also mentioned.Algorithm for binary word recognition suited to ultrafast nonlinear optics
http://dl-live.theiet.org/content/journals/10.1049/el_19930630
An algorithm is proposed for binary word recognition requiring only a single logical AND operation at the bit level. The AND recognition, represented by the scalar product of two optical fields, is fundamental to all ultrafast nonlinear optical phenomena.Binary encoded 2nd-differential spectrometry using UV-Vis spectral data and neural networks in the estimation of species type and concentration
http://dl-live.theiet.org/content/journals/10.1049/ip-smt_19970713
An approach to determining the type and concentration of a range of representative contaminants, chlorine, nitrate and ammonia in waste water, based on a three-stage scheme for processing data from ultraviolet and visible (UV-Vis) spectra, is described. In simulation in the laboratory, data for the study are derived from laboratory-based measurements of such spectra from mixtures of common chemical pollutants in water at levels around their legal limits and from mathematical models based on these measurements. Through the work, it is concluded that mathematical procedures alone, i.e. self-learning, are not currently effective, while classification based on a model for absorption spectra with prior knowledge of the expected chemistry in a particular water system under study, is more likely to be successful.Use of parametric modelling and statistical pattern recognition in detection of awareness during general anaesthesia
http://dl-live.theiet.org/content/journals/10.1049/ip-smt_19982324
Awareness is a rare but important complication of general anaesthesia. In its worst manifestation the patient is completely paralysed yet fully conscious and suffering the pain of the operative procedure. The sequelae from such an experience may be significant and lifelong. The paper describes a method, based on parametric modelling and statistical pattern recognition techniques, including neural networks, whereby awareness during general anaesthesia may be detected when present. Two systems are described, the first based solely on the use of the bispectrum, while the second makes use of both spectral and bispectral features. An evaluation on independent test sets shows that both systems have an average accuracy of > 80%, but the variation across individuals is less using the spectral–bispectral system (standard deviation of 16.4% compared with 20.5%). The spectral–bispectral system operates in near real time, requiring only 5 s of data to produce a new estimate of awareness. These estimates are obtained from the output of a trained neural network, which has as its input a set of features extracted from a single channel of electroencephalogram (EEG). The pre-processing of the data prior to input into the neural network is a critical component of the work, and it is here that parametric models have been extensively utilised. The spectral features are extracted from the EEG using a 1 s segment and a lattice filter as the primary model, while the bispectral features are extracted using a 5 s segment and a transversal filter as the underlying model.Nonstationary behaviour of partial discharge during discharge induced ageing of dielectrics
http://dl-live.theiet.org/content/journals/10.1049/ip-smt_19951435
Changes in the stochastic behaviour of pulsating partial discharge with time have been observed when an alternating voltage is applied to point-dielectric gaps in which the dielectric is a cast epoxy resin either with, or without, Al<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">2</sub>O<sub xmlns="http://pub2web.metastore.ingenta.com/ns/">3</sub> filler. The changes in discharge behaviour are shown, with the help of a Monte-Carlo simulation, to be consistent with discharge induced increases in the surface conductivity of the epoxy. This 'ageing' effect is shown to be accelerated as the frequency of applied voltage is increased from 50 to 800 Hz. The implications of the results on accelerated ageing tests and definition of partial discharge 'signatures' for possible pattern recognition are discussed. The relationship between average partial discharge current and partial discharge pulse height distributions is also discussed.Method of identifying individuals using VEP signals and neural network
http://dl-live.theiet.org/content/journals/10.1049/ip-smt_20040003
A method of identifying individuals using visual-evoked-potential (VEP) signals and neural network (NN) is proposed. In the approach, a backpropagation (BP) NN is trained to identify individuals using gamma-band (30-50 Hz) spectral power ratio of VEP signals extracted from 61 electrodes located on the scalp of the brain. The gamma-band spectral-power ratio is computed using a zero-phase Butterworth digital filter and Parseval's time-frequency equivalence theorem. NN classification gives an average of 99.06% across 400 test VEP patterns from 20 individuals using 10-fold cross-validation scheme. This shows promise for the approach to be developed further as a biometric identification system.Fuzzy uncertainty texture spectrum for texture analysis
http://dl-live.theiet.org/content/journals/10.1049/el_19950665
A new method using fuzzy uncertainty, which measures the uncertainty of the uniform surface in an image, is proposed for texture analysis. A grey-scale image can be transformed into a fuzzy image by the uncertainty definition. The distribution of the membership in a measured fuzzy image, denoted the fuzzy uncertainty texture spectrum (FUTS), is used as the texture feature for texture analysis.Fractal error for detecting man-made features in aerial images
http://dl-live.theiet.org/content/journals/10.1049/el_19940287
A technique is proposed for aiding photointerpreters in detecting man-made features in aerial reconnaissance images. The technique, which uses a metric called fractal error, is based on the observed propensity of natural image features to fit a fractional Brownian motion model. Man-made features usually do not fit this model well, and consequently the fractal error metric may be used as a discriminant function for detecting man-made scene features.