IET 3rd International Conference on Intelligent Signal Processing (ISP 2017)
Buy conference proceeding
- Location: London, UK
- Conference date: 4-5 Dec. 2017
- ISBN: 978-1-78561-707-2
- Conference number: CP731
- The following topics are dealt with: sequential Monte Carlo samplers; autonomous systems; 4D computer vision; Internet of Skills; 5G cellular network; and statistical methods.
19 items found
-
Performance analysis of UAV enabled disaster recovery network: a stochastic geometric framework based on matern cluster processes
- Author(s): A.M. Hayajneh ; S.A.R. Zaidi ; D.C. McLernon ; M. Ghogho
- + Show details - Hide details
-
p.
1 (6 .)
(1)
Drones will be employed by Facebook and Google for capacity off-loading in front/back hauling scenarios utilizing drone-empowered autonomous heterogeneous networks. But in another application, drone-based, post-disaster recovery of communication networks will also be of crucial importance in the design of future smart cities. So, in order to address the design issues of these latter networks, we present (from a stochastic geometric perspective) a comprehensive statistical framework for the spatial distribution of these hybrid user-centric drone/micro cellular networks. We introduce the novel idea of using a Stenien's cell (with variable radius) to model the region over which the drones will be distributed and the drones will effectively form a Matern cluster process (MCP) across the original network space. We then employ this newly developed framework to investigate the impact of changing several parameters on the performance of the new drone small-cell clustered networks (DSCCNs) and we develop appropriate closed-form expressions that model the performance (later validated via Monte Carlo simulations).
-
Appearance and Motion Information based Human Activity recognition
- Author(s): M. Al-Faris ; J. Chiverton ; L. Yang ; D. Ndzi
- + Show details - Hide details
-
p.
2 (6 .)
(1)
Activity recognition is an essential objective of a smart building system which responds to what is happening in a scene. In this paper, a view invariant activity recognition system is proposed to recognise human actions. Selection of applicable features is made and solutions are proposed to deal with probable challenges including differing views on actions and directionality issues. This paper explores a number of features that can be utilised in action recognition systems and chooses suitable features to mitigate the challenges properly. Motion History Image (MHI) based on historical appearance information is used in combination with local motion vectors which are computed through each iteration sequence of the MHI information using an optical flow algorithm. A multiview dataset (MuHaVi) and a single view dataset (Weizmann) are used to demonstrate and validate the proposed method. Our method, can detect a wide range of actions in multi-view scenarios and shows competitive performance in comparison with state-of-the-art action classification techniques.
-
Anomaly detection of aircraft engine in FDR (flight data recorder) data
- Author(s): Chang-Hun Lee ; Hyo-Sang Shin ; A. Tsourdos ; Z. Skaf
- + Show details - Hide details
-
p.
3 (6 .)
(1)
This paper deals with detection of anomalous behaviour of aircraft engines in FDR (flight data recorder) data to improve airline maintenance operations. To this end, each FDR data that records different flight patterns is first sampled at a fixed time interval starting at the take-off phase, in order to map each FDR data into comparable data space. Next, the parameters related to the aircraft engine are only selected from the sampled FDR data. In this analysis, the feature points are chosen as the mean value of each parameter within the sampling interval. For each FDR data, the feature vector is then formed by arranging all feature points. The proposed method compares the feature vectors of all FDR data and detects an FDR data in which the abnormal behaviour of the aircraft engine is recorded. The clustering algorithm called DBSCAN (density-based spatial clustering of applications with noise) is applied for this purpose. In this paper, the proposed method is tested using realistic FDR data provided by NASA's open database. The results indicate that the proposed method can be used to automatically identify an FDR data in which the abnormal behaviour of the aircraft engine is recorded from a large amount of FDR data. Accordingly, it can be utilized for a high-level diagnosis of engine failure in airline maintenance operations.
-
True colour retrieval from multiple illuminant scene's image
- Author(s): M.A. Hussain and A.S. Akbari
- + Show details - Hide details
-
p.
4 (5 .)
(1)
This paper presents an algorithm to retrieve the true colour of an image captured under multiple illuminant. The proposed method uses a histogram analysis and K-means++ clustering technique to split the input image into a number of segments. It then determines normalised average absolute difference (NAAD) for each resulting segment's colour component. If the NAAD of the segment's component is greater than an empirically determined threshold. It assumes that the segment does not represent a uniform colour area, hence the segment's colour component is selected to be used for image colour constancy adjustment. The initial colour balancing factor for each chosen segment's component is calculated using the Minkowski norm based on the principal that the average values of image colour components are achromatic. It finally calculates colour constancy adjustment factors for each image pixel by fusing the initial colour constancy factors of the chosen segments weighted by the normalised Euclidian distances of the pixel from the centroids of the selected segments. Experimental results using benchmark single and multiple illuminant image datasets, show that the proposed method's images subjectively exhibit highest colour constancy in the presence of multiple illuminant and also when image contains uniform colour areas.
-
Orientation Distribution and Mean Width Determination in micro X-Ray CT Images of Fibrous Materials
- Author(s): J. Chiverton ; A. Kao ; G. Tozzi ; M. Roldo
- + Show details - Hide details
-
p.
5 (6 .)
(1)
Fibrous nano-materials can be imaged using high resolution X-ray computer tomography (XCT). Some important parameters that are often manually estimated from this imaging data include the orientation distribution and the thickness of the fibres. Automation of this process is hampered by the close proximity of the fibres even with sub-micron voxel sizes. An automated sampling methodology has therefore been developed that can detect points in the imaging data where fibres are present and well separated. Polycaprolactone (PCL) electrospun fibrous material was prepared and imaged with a Zeiss Xradia Versa 510 Microtomography XCT. Automated measurements were then determined and further summarised using a single parameter measure.
-
Video compressive sensing for overlapped encoded frames
- Author(s): A. Matin and X. Wang
- + Show details - Hide details
-
p.
6 (6 .)
(1)
This paper investigates a particular application of compressive sensing (CS) for reconstruction of video data cubes in temporal multiplexing scheme (TM) and focuses on exploring the after acquisition reconstruction process. We propose a new implementation based on weighted 3D total variation algorithm with residual based updating parameters under the alternating direction method of multipliers (ADMM) framework for video CS applications. This method enables a fast data reconstruction with higher resolution compared to the other bench mark methods and benefits from iterative updates on optimisation parameters. We explore the reconstruction quality of the proposed method by comparing the results to some of the state of the art algorithms used in video CS and discuss the calculation time needed to process the data. Last section of this paper investigates the behaviour of the algorithm while handling the large amount of data and explores the potentials of proposed scheme to capture longer video sequences.
-
New approximation to design fractional order digital FIR differentiators
- Author(s): T. Bensouici ; A. Charef ; I. Assadi
- + Show details - Hide details
-
p.
7 (5 .)
(1)
In this paper an accurate and new digital approximation of the fractional-order differentiator (FOD) in the form of FIR filter is presented. This approach is based on power series expansion (PSE) of fractional order operators. First, an analog rational function approximation of the first order analog differentiator is given. The transformation of the analog domain to the discrete domain was made by using Bilinear transformation, which give us the digital version. Then, the digital FOD is obtained by taking fractional power of the ideal first order digital differentiator transfer function. Next, the PSE is applied to design fractional order digital FIR differentiator, with closed form formula. Finally, design examples are shown through the paper to illustrate the performance and the effectiveness of the proposed method.
-
The Optimization of Wideband Cyclostationary Feature Detector with Receiver Constraints
- Author(s): I.G. Anyim ; J. Chiverton ; M. Filip ; A. Tawfik
- + Show details - Hide details
-
p.
8 (6 .)
(1)
Cognitive radio system is a context-aware technology in communications. Spectrum awareness is an important function in the design of cognitive radio systems. It senses the presence or absence of primary users in the spectrum and declares the unoccupied channels for secondary users. Cyclostationary Feature Detection is about the detection of signals based on their features such as cyclic frequencies, symbol rates, carrier frequencies and modulation types. Detection can occur at very low signal to noise ratios.However performance degrading constraints such as cyclic and sampling clock offsets can occur at the receiver through the local oscillator frequency offsets, Doppler effects and jitter. We propose a multi-slot cyclostationary feature detector that reduces the effects of these constraints by optimizing for the number and size of each slot and fast Fourier transform. These slots and fast Fourier transforms are used to show the reduction of these offsets and the detection performance is compared for different scenarios with and without offsets.
-
Intermittently nonlinear analog filters for mitigation of technogenic interference in the process of analog-to-digital conversion
- Author(s): A.V. Nikitin and R.L. Davidchack
- + Show details - Hide details
-
p.
9 (7 .)
(1)
Since at any given frequency a linear filter affects both the noise and the signal of interest proportionally, when a linear filter is used to suppress the interference outside of the passband of interest the resulting signal quality is affected only by the total power and spectral composition, but not by the type of the amplitude distribution of the interfering signal. Thus a linear filter cannot improve the passband signal-to-noise ratio, regardless of the type of noise. On the other hand, a nonlinear filter has the ability to disproportionately affect signals with different temporal and/or amplitude structures, and it may reduce the spectral density of non-Gaussian (e.g. impulsive) interferences in the signal passband without significantly affecting the signal of interest. As a result, the signal quality can be improved in excess of that achievable by a linear filter. Such non-Gaussian (and, in particular, impulsive) noise can originate from a multitude of natural and technogenic (man-made) phenomena. The technogenic noise specifically is a ubiquitous and growing source of harmful interference affecting communication and data acquisition systems, and such noise may dominate over the thermal noise. While the non-Gaussian nature of technogenic noise provides an opportunity for its effective mitigation by nonlinear filtering, current state-of-the-art approaches employ such filtering in the digital domain, after analog-to-digital conversion. In the process of such conversion, the signal bandwidth is reduced, and the broadband non-Gaussian noise becomes more Gaussian-like. This substantially diminishes the effectiveness of the subsequent noise removal techniques. In this paper, we focus on impulsive noise mitigation, and propose to incorporate impulsive noise filtering of the analog input signal into nonlinear analog filters preceding an analog-to-digital converter (ADC). Such ADCs thus combine analog-to-digital conversion with analog rank filtering, enabling mitigation of various types of in-band non-Gaussian noise and interference, especially that of technogenic origin, including broadband impulsive interference. This can considerably increase quality of the acquired signal over that achievable by linear filtering in the presence of such interference. An important property of the presented approach is that, while being nonlinear in general, the proposed filters largely behave linearly. They exhibit nonlinear behaviour only intermittently, in response to noise outliers, thus avoiding the detrimental effects, such as instabilities and intermodulation distortions, often associated with nonlinear filtering.
-
Partial-Update Adaptive Filters for Event-Related Potentials Denoising
- Author(s): M. Boudiaf ; M. Benkherrat ; M.A. Boudiaf
- + Show details - Hide details
-
p.
10 (6 .)
(1)
In the present study, the adaptive Fourier linear combiner is used to denoise the single-trial event-related potential signals. The Least Mean Square algorithm, the Normalized Least Mean Square algorithm and the Recursive Least Square algorithm are compared with different approaches to partial coefficient updates: periodic partial updates, sequential partial updates, stochastic partial updates, M-max updates and selective partial updates. The different approaches are compared in term of convergence speed, tracking capability and computational complexity. The methods are also validated on a real human P300 recording.
-
Parallelising Particle Filters with Deterministic Runtime on Distributed Memory Systems
- Author(s): A. Varsi ; L. Kekempanos ; J. Thiyagalingam ; S. Maskell
- + Show details - Hide details
-
p.
11 (10 .)
(1)
Particle Filters are parallelisable. However, the resampling step, which is necessary to handle the degeneracy problem, is non-trivial to parallelise. A textbook implementation of this step achieves O (N) time complexity with a very low constant but deriving an efficient parallel implementation of the same is non-trivial. Several alternative solutions have been proposed. A state-of-the-art parallelisable algorithm has O (log2N)2 time complexity and has been implemented in the Big data MapReduce context. However, portability of this algorithm is difficult to achieve, especially on distributed memory systems. In paper, we reformulate this algorithm for the distributed memory setup and demonstrate that the same algorithm can be almost twice as fast as the exiting state-of-theart algorithm for up to 128 cores. Furthermore, we also show that our algorithm is 20 times faster than the serial version of the algorithm.
-
Approximate message passing for underdetermined audio source separation
- Author(s): T. Iqbal and Wenwu Wang
- + Show details - Hide details
-
p.
12 (5 .)
(1)
Approximate message passing (AMP) algorithms have shown great promise in sparse signal reconstruction due to their low computational requirements and fast convergence to an exact solution. Moreover, they provide a probabilistic framework that is often more intuitive than alternatives such as convex optimisation. In this paper, AMP is used for audio source separation from underdetermined instantaneous mixtures. In the time-frequency domain, it is typical to assume a priori that the sources are sparse, so we solve the corresponding sparse linear inverse problem using AMP. We present a block-based approach that uses AMP to process multiple time-frequency points simultaneously. Two algorithms known as AMP and vector AMP (VAMP) are evaluated in particular. Results show that they are promising in terms of artefact suppression.
-
Regression analysis for paths inference in a novel Proton CT system
- Author(s): Liyun Gong ; Miao Yu ; Xujiong Ye ; N. Allinson
- + Show details - Hide details
-
p.
13 (4 .)
(1)
In this work, we analyse the proton paths inference for the construction of CT imagery based on a new proton CT proton system, which can record multiple proton paths/residual energies. Based on the recorded paths of multiple protons, every proton path is inferred. The inferred proton paths can then be used for the residual energies detection and CT imagery construction for analyzing a specific tissue. Different regression methods (linear regression and Gaussian process regression models) are exploited for the path inference of every proton in this work. The studies on a recorded proton trajectories dataset show that the Gaussian process regression method achieves better accuracies for the path inference, from both path assignment accuracy and root mean square errors (RMSEs) studies.
-
Automatic Detection of Speech Disorder in Dysarthria using Extended Speech Feature Extraction and Neural Networks Classification
- Author(s): T.B. Ijitona ; J.J. Soraghan ; A. Lowit ; G. Di-Caterina ; H. Yue
- + Show details - Hide details
-
p.
15 (6 .)
(1)
This paper presents an automatic detection of Dysarthria, a motor speech disorder, using extended speech features called Centroid Formants. Centroid Formants are the weighted averages of the formants extracted from a speech signal. This involves extraction of the first four formants of a speech signal and averaging their weighted values. The weights are determined by the peak energies of the bands of frequency resonance, formants. The resulting weighted averages are called the Centroid Formants. In our proposed methodology, these centroid formants are used to automatically detect Dysarthric speech using neural network classification technique. The experimental results recorded after testing this algorithm are presented. The experimental data consists of 200 speech samples from 10 Dysarthric speakers and 200 speech samples from 10 age-matched healthy speakers. The experimental results show a high performance using neural networks classification. A possible future research related to this work is the use of these extended features in speaker identification and recognition of disordered speech.
-
Learning to approximate computing at run-time
- Author(s): P. Garcia ; M. Emambakhsh ; A. Wallace
- + Show details - Hide details
-
p.
16 (6 .)
(1)
Intelligent sensor/signal processing systems are increasingly constrained by tight power budgets, especially when deployed in mobile/remote environments. Approximate computing is the process of adaptively compromising the accuracy of a system's output in order to obtain higher performance for other metrics, such as power consumption or memory usage, for applications resilient to inaccurate computations. It is, however, usually statically implemented, based on heuristics and testing loops, which prevents switching between different approximations at run-time. This limits approximation versatility and results in under- or over-approximated systems for the specific input data, causing excessive power usage and/or insufficient accuracy, respectively. To avoid these issues, this paper proposes a new approximate computing approach by introducing a supervisor block embedding prior knowledge about runtime data. The target system (i.e., signal processing pipeline) is implemented with configurable levels and types of approximations [1]. Data processed by the target system is analysed by the supervisor and the approximation is updated dynamically, by using prior knowledge to establish a confidence measure on the accuracy of the computed results. Moreover, by iteratively evaluating the output, the supervisor block can learn and subsequently update tunable parameters, to improve the quality of the results. We detail and evaluate this approach for tracking problem in computer vision. Results show our approach yields promising trade-offs between accuracy and power consumption, achieving 2.54% energy saving for our case study.
-
Channel Variability Synthesis in i-vector Speaker Recognition
- Author(s): A.I. Ahmed ; J. Chiverton ; D. Ndzi ; V. Becerra
- + Show details - Hide details
-
p.
17 (6 .)
(1)
In this paper, we are tackling a practical problem which can be faced when establishing an i-vector speaker recognition system with limited resources. This addresses the problem of lack of development data of multiple recordings for each speaker. When we only have one recording for each speaker in the development set, phonetic variability can be simply synthesised by dividing the recordings if they are of sufficient length. For channel variability, we pass the recordings through a Gaussian channel to produce another set of recordings, referred to here as Gaussian version recordings. The proposed method for channel variability synthesis produces total relative improvements in EER of 5%.
-
Effects of Acoustic Features Modifications on the Perception of Dysarthric Speech - Preliminary Study (Pitch, Intensity and Duration Modifications)
- Author(s): T.B. Ijitona ; J.J. Soraghan ; A. Lowit ; G. Di-Caterina ; H. Yue
- + Show details - Hide details
-
p.
18 (6 .)
(1)
Marking stress is important in conveying meaning and drawing listener's attention to specific parts of a message. Extensive research has shown that healthy speakers mark stress using three main acoustic cues; pitch, intensity, and duration. The relationship between acoustic and perception cues is vital in the development of a computer-based tool that aids the therapists in providing effective treatment to people with Dysarthria. It is, therefore, important to investigate the acoustic cues deficiency in dysarthric speech and the potential compensatory techniques needed for effective treatment. In this paper, the relationship between acoustic and perceptive cues in dysarthric speech are investigated. This is achieved by modifying stress marked sentences from 10 speakers with Ataxic dysarthria. Each speaker produced 30 sentences using the 10 SubjectVerb-Object-Adjective (SVOA) structured sentences across three stress conditions. These stress conditions are stress on the initial (S), medial (O) and final (A) target words respectively. To effectively measure the deficiencies in Dysarthria speech, the acoustic features (pitch, intensity, and duration) are modified incrementally. The paper presents the techniques involved in the modification of these acoustic features. The effects of these modifications are analysed based on steps of 25% increments in pitch, intensity and duration. For robustness and validation, 50 untrained listeners participated in the listening experiment. The results and the relationship between acoustic modifications (what is measured) and perception (what is heard) in Dysarthric speech are discussed.
-
A Mixed resolution based High Efficiency Video Codec (HEVC)
- Author(s): B. Mallik ; A.S. Akbari ; M.A. Hussain ; A.L. Kor
- + Show details - Hide details
-
p.
19 (5 .)
(1)
Development of efficient video codecs for low bitrate video transmission with higher compression efficiency has been an active area of research over past years and various techniques have been proposed to fulfil this need. In this paper, a mixed resolution based video codec for low bitrate transmission within the standard HEVC codec's framework is proposed. A spatial resolution scaling type of mixed resolution coding model for monoscopic videos using HEVC codec is presented. The proposed mixed-resolution structure and reference frames structure simplifies the implementation of a mixed-resolution based HEVC codec that can code video frames with different resolutions. In order to evaluate the performance of the proposed codec; three 4:2:0 format test video sequences, namely “Cactus”, "KristenAndSara" and “ParkScene”, were selected and coded using the proposed codec and the standard HEVC codec. Experimental results show that the proposed mixed resolution based HEVC codec gives a significantly higher coding performance to that of the standard HEVC codec at low bitrates.
-
Optimized Stockwell transform for the segmentation of ECG signals
- Author(s): Z. Bouguila ; A. Moukadem ; A. Dieterlen
- + Show details - Hide details
-
p.
14 (4 .)
(1)
In this study, an optimized Stockwell transform is applied on ECG signals. The optimization of time-frequency representation helps to separate better the different complex of ECG signal. The envelope of the signal is extracted by applying non linear transform on the optimized time-frequency matrix and a peak detector algorithm is applied to detect the different ECG complex. The proposed method is evaluated on synthetic and real ECG signals for healthy adult's volunteers. The robustness against noise of the proposed method is also tested and discussed.
