Engineering Science, School of

Receive updates for this collection

De-sketching

Peer reviewed: 
Yes, item is peer reviewed.
Date created: 
2018-06-22
Abstract: 

Many software applications exist for plotting graphs of mathematical functions, yet there are none (to our knowledge) that perform the inverse operation - estimating mathematical expressions from graphs. Since plotting graphs (especially by hand) is often referred to as "sketching," we refer to the inverse operation as "de-sketching." As the number of mathematical expressions that approximate a given curve can be quite large, in this demo we restrict our attention to polynomials, and present a deep model that performs de-sketching by finding the best second-degree polynomial to fit the curve in the input image. Currently, our trained model is able to provide reasonably accurate estimates of polynomial coefficients for both synthetically-generated and hand-drawn curves.

Document type: 
Conference presentation
File(s): 

DFTS: Deep Feature Transmission Simulator

Peer reviewed: 
Yes, item is peer reviewed.
Date created: 
2018-06-22
Abstract: 

Collaborative intelligence is a deployment paradigm for deep AI models where some of the layers run on the mobile terminal or network edge, while others run in the cloud. In this scenario, features computed in the model need to be transferred between the edge and the cloud over an imperfect channel. Here we present a simulator to help study the effects of imperfect packet-based transmission of deep features. Our simulator is implemented in Keras and allows users to study the effects of both lossy packet transmission and quantization on the accuracy.

Document type: 
Conference presentation
File(s): 

An Improved Theoretical Process-Zone Model for Delayed Hydride Cracking Initiation at a Blunt V-Notch

Peer reviewed: 
Yes, item is peer reviewed.
Date created: 
2018-01-30
Abstract: 

Delayed hydride cracking (DHC) is an important concern for pressure tubes used in nuclear reactors.  In this paper, an improved analytical process-zone model is developed based on the deformation fracture criteria. A V-notch with rounded root, which is widely adopted in mechanical testing of DHC, is considered and the proposed model includes the effect of both notch angle and tip radius. Comparisons with experiments show that the proposed model has a prediction accuracy closer to the current engineering process-zone model but with slightly less conservatism. The model is extended to account for plasticity and constraint effects at the flaw tip by introducing an empirical factor that depends on key material and geometric parameters.

Document type: 
Article

Atomic-Scale Finite Element Modelling of Mechanical Behaviour of Graphene Nanoribbons

Peer reviewed: 
Yes, item is peer reviewed.
Date created: 
2018-02-09
Abstract: 

Experimental characterization of Graphene NanoRibbons (GNRs) is still an expensive task and computational simulations are therefore seen a practical option to study the properties and mechanical response of GNRs. Design of GNR in various nanotechnology devices can be approached through molecular dynamics simulations. This study demonstrates that the Atomic–scale Finite Element Method (AFEM) based on the second generation REBO potential is an efficient and accurate alternative to the molecular dynamics simulation of GNRs. Special atomic finite elements are proposed to model graphene edges. Extensive comparisons are presented with MD solutions to establish the accuracy of AFEM. It is also shown that the Tersoff potential is not accurate for GNR modeling. The study demonstrates the influence of chirality and size on design parameters such as tensile strength and stiffness. A GNR is stronger and stiffer in the zigzag direction compared to the armchair direction. Armchair GNRs shows a minor dependence of tensile strength and elastic modulus on size whereas in the case of zigzag GNRs both modulus and strength show a significant size dependency. The size-dependency trend noted in the present study is different from the previously reported MD solutions for GNRs but qualitatively agrees with experimental results. Based on the present study, AFEM can be considered a highly efficient computational tool for analysis and design of GNRs.

Document type: 
Article

Spontaneous Blinks Activate the Precuneus: Characterizing Blink-Related Oscillations Using Magnetoencephalography

Peer reviewed: 
Yes, item is peer reviewed.
Date created: 
2017-09-26
Abstract: 

Spontaneous blinking occurs 15–20 times per minute. Although blinking has often been associated with its physiological role of corneal lubrication, there is now increasing behavioral evidence suggesting that blinks are also modulated by cognitive processes such as attention and information processing. Recent low-density electroencephalography (EEG) studies have reported so-called blink-related oscillations (BROs) associated with spontaneous blinking at rest. Delta-band (0.5–4 Hz) BROs are thought to originate from the precuneus region involved in environmental monitoring and awareness, with potential clinical utility in evaluation of disorders of consciousness. However, the neural mechanisms of BROs have not been elucidated. Using magnetoencephalography (MEG), we characterized delta-band BROs in 36 healthy individuals while controlling for background brain activity. Results showed that, compared to pre-blink baseline, delta-band BROs resulted in increased global field power (p < 0.001) and time-frequency spectral power (p < 0.05) at the sensor level, peaking at ∼250 ms post-blink maximum. Source localization showed that spontaneous blinks activated the bilateral precuneus (p < 0.05 FWE), and source activity within the precuneus was also consistent with sensor-space results. Crucially, these effects were only observed in the blink condition and were absent in the control condition, demonstrating that results were due to spontaneous blinks rather than as part of the inherent brain activity. The current study represents the first MEG examination of BROs. Our findings suggest that spontaneous blinks activate the precuneus regions consistent with environmental monitoring and awareness, and provide important neuroimaging support for the cognitive role of spontaneous blinks.

Document type: 
Article
File(s): 

Joint Source-Channel Coding of JPEG 2000 Image Transmission Over Two-Way Multi-Relay Networks

Peer reviewed: 
Yes, item is peer reviewed.
Date created: 
2017-04
Abstract: 

In this paper, we develop a two-way multi-relay scheme for JPEG 2000 image transmission. We adopt a modified time-division broadcast (TDBC) cooperative protocol, and derive its power allocation and relay selection under a fairness constraint. The symbol error probability of the optimal system configuration is then derived. After that, a joint source-channel coding (JSCC) problem is formulated to find the optimal number of JPEG 2000 quality layers for the image and the number of channel coding packets for each JPEG 2000 codeblock that can minimize the reconstructed image distortion for the two users, subject to a rate constraint. Two fast algorithms based on dynamic programming (DP) and branch and bound (BB) are then developed. Simulation demonstrates that the proposed JSCC scheme achieves better performance and lower complexity than other similar transmission systems.

Document type: 
Article
File(s): 

Load Disaggregation Based on Aided Linear Integer Programming

Peer reviewed: 
Yes, item is peer reviewed.
Date created: 
2016
Abstract: 

Load disaggregation based on aided linear integer programming (ALIP) is proposed. We start with a conventional linear integer programming (IP) based disaggregation and enhance it in several ways. The enhancements include additional constraints, correction based on a state diagram, median filtering, and linear programming-based refinement. With the aid of these enhancements, the performance of IP-based disaggregation is significantly improved. The proposed ALIP system relies only on the instantaneous load samples instead of waveform signatures, and hence works well on low-frequency data. Experimental results show that the proposed ALIP system performs better than conventional IP-based load disaggregation.

Document type: 
Article
File(s): 

Online MoCap Data Coding with Bit Allocation, Rate Control, and Motion-Adaptive Post-Processing

Peer reviewed: 
Yes, item is peer reviewed.
Date created: 
2017-01
Abstract: 

With the advancements in methods for capturing 3D object motion, motion capture (MoCap) data are starting to be used beyond their traditional realm of animation and gaming in areas such as the arts, rehabilitation, automotive industry, remote interactions, and so on. As the amount of MoCap data increases, compression becomes crucial for further expansion and adoption of these technologies. In this paper, we extend our previous work on low-delay MoCap data compression by introducing two improvements. The first improvement is the bit allocation to long-term and short-term reference MoCap frames, which provides a 10-15% reduction in coded bitrate at the same quality. The second improvement is the post-processing in the form of motion-adaptive temporal low-pass filtering, which is able to provide another 9-13%savings in the bitrate. The experimental results also indicate that the proposed online MoCap codec is competitive with several state-of-the-art offline codecs. Overall, the proposed techniques integrate into a highly effective online MoCap codec that is suitable for low-delay applications, whose implementation is provided alongside this paper to aid further research in the field.

Document type: 
Article
File(s): 

Label-Free Density Measurements of Radial Peripapillary Capillaries in the Human Retina

Peer reviewed: 
Yes, item is peer reviewed.
Date created: 
2015
Abstract: 

Radial peripapillary capillaries (RPCs) comprise a unique network of capillary beds within the retinal nerve fibre layer (RNFL) and play a critical role in satisfying the nutritional requirements of retinal ganglion cell (RGC) axons. Understanding the topographical and morphological characteristics of these networks through in vivo techniques may improve our understanding about the role of RPCs in RGC axonal health and disease. This study utilizes a novel, non-invasive and label-free optical imaging technique, speckle variance optical coherence tomography (svOCT), for quantitatively studying RPC networks in the human retina. Six different retinal eccentricities from 16 healthy eyes were imaged using svOCT. The same eccentricities were histologically imaged in 9 healthy donor eyes with a confocal scanning laser microscope. Donor eyes were subject to perfusion-based labeling techniques prior to retinal dissection, flat mounting and visualization with the microscope. Capillary density and diameter measurements from each eccentricity in svOCT and histological images were compared. Data from svOCT images were also analysed to determine if there was a correlation between RNFL thickness and RPC density. The results are as follows: (1) The morphological characteristics of RPC networks on svOCT images are comparable to histological images; (2) With the exception of the nasal peripapillary region, there were no significant differences in RPC density measurements between svOCT and histological images; (3) Capillary diameter measurements were significantly greater in svOCT images compared to histology; (4) There is a positive correlation between RPC density and RNFL thickness. The findings in this study suggest that svOCT is a reliable modality for analyzing RPC networks in the human retina. It may therefore be a valuable tool for aiding our understanding about vasculogenic mechanisms that are involved in RGC axonopathies. Further work is required to explore the reason for some of the quantitative differences between svOCT and histology.

Document type: 
Article
File(s): 

EEG Classification of Different Imaginary Movements within the Same Limb

Peer reviewed: 
Yes, item is peer reviewed.
Date created: 
2015
Abstract: 

The task of discriminating the motor imagery of different movements within the same limb using electroencephalography (EEG) signals is challenging because these imaginary movements have close spatial representations on the motor cortex area. There is, however, a pressing need to succeed in this task. The reason is that the ability to classify different same-limb imaginary movements could increase the number of control dimensions of a brain-computer interface (BCI). In this paper, we propose a 3-class BCI system that discriminates EEG signals corresponding to rest, imaginary grasp movements, and imaginary elbow movements. Besides, the differences between simple motor imagery and goal-oriented motor imagery in terms of their topographical distributions and classification accuracies are also being investigated. To the best of our knowledge, both problems have not been explored in the literature. Based on the EEG data recorded from 12 able-bodied individuals, we have demonstrated that same-limb motor imagery classification is possible. For the binary classification of imaginary grasp and elbow (goal-oriented) movements, the average accuracy achieved is 66.9%. For the 3-class problem of discriminating rest against imaginary grasp and elbow movements, the average classification accuracy achieved is 60.7%, which is greater than the random classification accuracy of 33.3%. Our results also show that goal-oriented imaginary elbow movements lead to a better classification performance compared to simple imaginary elbow movements. This proposed BCI system could potentially be used in controlling a robotic rehabilitation system, which can assist stroke patients in performing task-specific exercises.

Document type: 
Article
File(s):