Gordon Morison

A reconfigurable real-time morphological system for augmented vision

There is a significant number of visually impaired individuals who suffer sensitivity loss to high spatial frequencies, for whom current optical devices are limited in degree of visual aid and practical application. Digital image and video processing offers a variety of effective visual enhancement methods that can be utilised to obtain a practical augmented vision head-mounted display device. The high spatial frequencies of an image can be extracted by edge detection techniques and overlaid on top of the original image to improve visual perception among the visually impaired. Augmented visual aid devices require highly user-customisable algorithm designs for subjective configuration per task, where current digital image processing visual aids offer very little user-configurable options. This paper presents a highly user-reconfigurable morphological edge enhancement system on field-programmable gate array, where the morphological, internal and external edge gradients can be selected from the presented architecture with specified edge thickness and magnitude. In addition, the morphology architecture supports reconfigurable shape structuring elements and configurable morphological operations. The proposed morphology-based visual enhancement system introduces a high degree of user flexibility in addition to meeting real-time constraints capable of obtaining 93 fps for high-definition image resolution.

Extended fast compressive tracking with weighted multi-frame template matching for fast motion tracking

State of the art tracking algorithm significantly extended.Multi-frame template matching greatly improves tracking accuracy.Tracking of fast moving targets and motion blur improved.Real-time operation in excess of 120 frames per secondAlgorithm performs strongly against state of the art. The work presented in this letter extends upon state of the art visual object tracking algorithms, the Real-time Compressive Tracker and the Fast Compressive Tracker, increasing the overall tracking accuracy at a minimal computational cost and reduction in frame rate. A template matching processing stage is incorporated in order to increase the robustness of the algorithm while maintaining a frame rate well within the requirements for real time operation. We utilise a weighted multi-frame similarity metric, template matching a bank of the top classifier outputs against the ground truth bounding box and a recently stored target bounding box to select the appropriate target location in the following frame. Unlike the original algorithm, the proposed method utilises more of the available data to make more informed tracking decisions than purely using the highest classifier output. Multiple similarity metrics have been employed in the template matching stage to compare their performance on a range of commonly used publicly available image sequences. The extended algorithm clearly demonstrated an increase in the overall performance while maintaining a high frame-rate.

Forensic analysis of Kik messenger on iOS devices

Instant messaging applications continue to grow in popularity as a means of communicating and sharing multimedia files. The information contained within these applications can prove invaluable to law enforcement in the investigation of crimes.Kik messenger is a recently introduced instant messaging application that has become very popular in a short period of time, especially among young users. The novelty of Kik means that there has been little forensic examination conducted on this application.This study addresses this issue by investigating Kik messenger on Apple iOS devices. The goal was to locate and document artefacts created or modified by Kik messenger on devices installed with the latest version of iOS, as well as in iTunes backup files. Once achieved, the secondary goal was to analyse the artefacts to decode and interpret their meaning and by doing so, be able to answer the typical questions faced by forensic investigators.A detailed description of artefacts created or modified by Kik messenger is provided. Results from experiments showed that deleted images are not only recoverable from the device, but can also be located and downloaded from Kik servers. A process to link data from multiple database tables producing accurate chat histories is explained. These outcomes can be used by law enforcement to investigate crimes and by software developers to create tools to recover evidence.

An implementation focused approach to teaching image processing and machine vision - from theory to beagleboard

The explosion of multimedia applications within embedded devices has ensured that Image Processing and Machine Vision has now become a mainstream subject within most Computer Science and Electronic Engineering curricula. Yet often there exists a disconnection between the rapid prototyping tools that are taught within the laboratory to demonstrate concepts and those that are used for actual deployment in a stand-alone product. This calls for an approach whereby students are exposed to multiple levels of abstraction, in order to align the skill sets of our students with the requirements and expectations of industry. This paper describes the development of a senior level undergraduate course that introduces machine vision and image processing algorithms and implementation topics within the larger context of embedded computing. The key focus is that the student appreciates the theoretical concepts but is also capable of implementing them on embedded processors for prototyping or production.

Selective Sampling Importance Resampling Particle Filter Tracking With Multibag Subspace Restoration.

The focus of this paper is a novel object tracking algorithm which combines an incrementally updated subspace-based appearance model, reconstruction error likelihood function and a two stage selective sampling importance resampling particle filter with motion estimation through autoregressive filtering techniques. The primary contribution of this paper is the use of multiple bags of subspaces with which we aim to tackle the issue of appearance model update. The use of a multibag approach allows our algorithm to revert to a previously successful appearance model in the event that the primary model fails. The aim of this is to eliminate tracker drift by undoing updates to the model that lead to error accumulation and to redetect targets after periods of occlusion by removing the subspace updates carried out during the period of occlusion. We compare our algorithm with several state-of-the-art methods and test on a range of challenging, publicly available image sequences. Our findings indicate a significant robustness to drift and occlusion as a result of our multibag approach and results show that our algorithm competes well with current state-of-the-art algorithms.

Hardware Threading Techniques for Multi-Threaded MPSoCs

Adapting software applications to embedded Multiprocessor System on Chips (MPSoCs) typically follows multithreaded design flows. To take advantage of the hardware customisations possible with MPSoCs, HardWare Threads (HWTs) can be used to increase application concurrency and throughput by forking between software and hardware execution. This paper describes how an application can be tailored to use HWTs. Using an applications Task Flow Graph and Kahn Process Networks to model software interactions with HWTs, two scheduling techniques for HWT interaction with software are presented and analysed. The scheduling techniques are evaluated based on system performance and resource consumption with a popular image processing algorithm, where performance increases of up to 3.6x were measured compared to standard implementations.

Multiscale permutation entropy analysis of the EEG in early stage alzheimer's patients

Alzheimers Disease (AD) is a neurodegenerative disorder associated with a progressive loss of cognitive function. Early identification of AD, when symptoms are mild, can be difficult. Therefore, the development of clinically useful measures are necessary to improve diagnosis of the disease and to allow for early clinical intervention, as well as to aid in drug development. The aim of this study is to analyse the electroencephalography (EEG) of patients with mild AD while they were engaged in a memory task, and to contrast these results with those from cognitively healthy control subjects. We introduce a novel application of the Multiscale Permutation Entropy (MPE) analysis to the EEG signal of patients and controls during task execution, which allows us to compare the complexity of the underlying brain signals at multiple temporal scales. These complexity results are then correlated with cognitive behavioral measures to evaluate the correspondence between complexity and cognitive performance.

Custom memory architecture for multi-core implementation of face detection algorithm

Here we present our multi-core architectures for object detection. We move away from the traditional architecture of Multi-Processors (MPs) by using cacheable accesses to main memory to create atomic cores and utilising local memory for all program data. Main memory is partitioned through software into dedicated data regions to allow atomic accesses by cores, without the need for synchronisation primitives. In doing this, we demonstrate how multi-threading techniques such as Interleaved Task Reordering (ITR) can be utilised to balance the processing loads on available cores. We implement and test up to 7 soft-cores with the Viola Jones face detection algorithm and achieve a performance increase of up to 9.14x with a 100% detection rate: surpassing the theoretical performance increase of multi-core processors for all designs and test images. Furthermore, we surpass the performance increases of multi-core implementations from the literature, thus proving our custom designs to be a more viable solution for multi-core object detection applications. Finally, resource and power consumption estimates indicate our designs to be suitable for embedded systems deployment.

Classification of partial discharge EMI conditions using permutation entropy-based features

In this paper we investigate the application of feature extraction and machine learning techniques to fault identification in power systems. Specifically we implement the novel application of Permutation Entropy-based measures known as Weighted Permutation and Dispersion Entropy to field Electro-Magnetic Interference (EMI) signals for classification of discharge sources, also called conditions, such as partial discharge, arcing and corona which arise from various assets of different power sites. This work introduces two main contributions: the application of entropy measures in condition monitoring and the classification of real field EMI captured signals. The two simple and low dimension features are fed to a Multi-Class Support Vector Machine for the classification of different discharge sources contained in the EMI signals. Classification was performed to distinguish between the conditions observed within each site and between all sites. Results demonstrate that the proposed approach separated and identified the discharge sources successfully.

An extended Real-Time Compressive Tracking Method using weighted multi-frame Cosine Similarity Metric

This paper presents an extended algorithm for Real-time Compressive Tracking using Cosine Similarity Metric for object tracking. The method utilises a weighted multi-frame cosine similarity metric with the ground truth bounding box and a recently computed target bounding box. In comparison to the original algorithm it is capable of handling fast motion with a greater degree of accuracy. The proposed algorithm has been benchmarked on a desktop computer and subsequently implemented on a Texas Instruments ARM based DM3730 Beagleboard-xM. The proposed algorithm demonstrates a significant performance increase in fast motion video sequences. In addition, the low computational complexity of the algorithm makes it well suited for embedded applications.