Cohort 3 (2021)
Machine learning models in ultrasound tongue imaging for the detection of children's speech disorders
I’m Saja Al Ani, a PhD student in the Future Ultrasonic Engineering (FUSE) CDT. I have completed a Bachelor of Science in Information Technology, followed by an M.Sc. degree in E-learning Technology from the University of Hertfordshire. Since joining the EPSRC Centre of Doctoral Training FUSE in the year 2021, I have gained a great deal of knowledge and experience in ultrasonic engineering. I developed an understanding of the latest developments in ultrasonic technology and their impact on a variety of applications during this comprehensive programme.
Currently in Speech and Language Therapy (SLT), technological support is sparse. Assessment of speech disorders is particularly time consuming and suffers from a lack of technological solutions. Normally speech assessment involves listening to the child and writing down what they say. This approach can miss important subtleties in the way children speak. For example, a child may try to say “key” and it may be heard as “tea”. This leads the SLT to believe the child cannot tell the difference between t and k and select a therapy designed to tackle this. However, using medical ultrasound to image tongue movements reveals that in many cases children are producing imperceptible errors. This is particularly the case for children with cleft lip and palate who produces a wide variety of unusual speech errors. However, ultrasound analysis is a time consuming task which can only be carried out by a speech scientist with specialist training. My project aims to use machine learning approaches for method for classifying tongue shapes in children with cleft lip and palate. This has the potential to be useful both for assessment and for measuring progress in intervention.
Primary Supervisor (University of Glasgow): Dr. Ahmed Zoha
Secondary Supervisor (University of Strathclyde): Dr. Joanne Cleland
Project External Partner: NHS Greater Glasgow & Clyde
Wearable Ultrasound Sensors for Sonomyography in Robotic Prosthesis and Augmentation
I’m Priyanka Dhiwa, a graduate from Amity University India with a Bachelor and Master of Technology degree (Distinction) in Nanoscience and Nanotechnology. Following a year internship at National Physical Laboratory (NPL), India, my Master’s project involves “Electrocaloric effect of ferroelectric polycrystalline ceramics for solid-state Refrigerator”.
SonoMyoGraphy (SMG) is the use of an ultrasonic transducer (10 MHz) to measure muscle activity. It is an alternative to electromyography (EMG) offering non-invasive and higher spatial resolution. SMG is a promising technique for muscle condition diagnosis, rehabilitation engineering, prosthesis control, and augmentation. The difficulty of distinguishing between the muscle action of the primary muscle and that of neighbouring muscles and the inefficiency of scanning deeper muscles are two of the most significant drawbacks of employing EMG.
My project aims to demonstrate the potential of sonomyography as an imaging technique for the assessment of muscle activity which could lead to more efficient and accurate diagnosis of myopathies and damage to the peripheral nervous system. In order to do this, the first step will be to decide the best material for such a task. Since my project will be focused around the use of microultrasound through a flexible piezoelectric sensor, the main materials found to be of interest have been lead zirconate titanate (PZT), polyvinylidene fluoride (PVDF) and aluminium nitride (AIN) since these have shown adequate results as demonstrated in the state of the art. It is worth noting that microultrasound, also known as high-frequency ultrasound, has been chosen because of the enhanced spatial resolution that it offers, although this could pose an obstacle due to the imaging depth decreasing. Such an obstacle could be potentially avoided by making this an implantable device, which will also be investigated during the project.
Once the most suitable material is chosen, the device will go through different design phases: single electrode, array, acoustic matching layer and acoustic lens, in order to optimise the design proposed. Finally, my project will conclude by investigating the modulation on the muscles caused by the microultrasound patch design. Such methodology has been previously shown in the literature.
Primary Supervisor (University of Glasgow): Prof. Sandy Cochran
Secondary Supervisor (University of Strathclyde): Prof. James Windmill
Project External Partner: Neuranics
Development of Ultrasound based device for intracranial ventriculostomy
I graduated with a MD diploma from Medical University of Lublin, English Division, in 2016 and subsequently worked as a doctor, initially in Poland as an intern and then as a Foundation Trainee in Glasgow for NHS Greater Glasgow and Clyde. After a further year as a Clinical Fellow in Orthopaedics, I studied Biomedical Engineering at Strathclyde University, with my thesis being ultrasound-based. Currently I’m working less than full time as a GP trainee and devoting the rest of my time to FUSE. I became interested in FUSE as I wanted to combine both my medical and engineering backgrounds and always had special interest in ultrasonics from the first time I held an ultrasound transducer as a medical student. My special interest areas are medical ultrasound including detection of pathologies, elastography and HIFU.
The aim of my project is to create an ultrasound-guided drainage tube that is designed specifically for the intracranial ventriculostomy procedure.
The envisaged device will not only make the procedure much easier but might facilitate it being done bedside, to make it much more time and economically efficient and easier for the patient. Other usage ideas are in place which can reduce the need for further imaging for those patients with an EVD in place in case the tube is obstructed, which is not uncommon. The ventriculostomy procedure is part of daily neurosurgical practice. According to a UK study, there were nearly 500 EVDs inserted in a specific six-month period. This does not include permanent shunt (drain) insertion, totalling a further 3000 shunts per year, approximately, according to the UK shunt registry. The global EVD market size was valued at USD 5.3 billion in 2019 and is expected to expand at a compound annual growth rate (CAGR) of 7.6% between 2020 and 2027.
There is evidence supporting the use of image guidance during tube insertion to improve accuracy and avoid misplaced catheters. However, neurosurgeons might be reluctant to use image guidance to avoid time delays or because of their lack of experience with using ultrasound. Designing an easy-to-use, specific and efficient ultrasound imaging tool with a seamless user experience would therefore be of added value to the neurosurgical practice. Current activity can be divided into image-guided solution providers and tube (shunt) manufacturers. Neuro-navigation manufacturers includes Stealth and Brainlab and ultrasound manufacturers such as BK Medical. Shunt manufacturers are also relevant, if the new design is configures as an ultrasound imaging device integrated with the drainage tube. There are many relevant manufacturers including but not limited to Medtronic and Stryker.
Primary Supervisor (University of Glasgow): Prof. Sandy Cochran
Secondary Supervisor (University of Strathclyde): Prof. James Windmill
Project External Partner: Astroin
Automated Inspection of Challenging Components using Novel Kelpie Designs
I achieved distinction in MSc Mechanical Engineering (Design) from Glasgow Caledonian University. I have a BEng (Hons) in Mechanical Engineering with thorough hands-on experience in design, manufacturing and testing within the space sector. By undertaking a successful 9-month work placement at Airbus Defence and Space in Germany, along with an Honours Project in partnership with the company, I developed both my technical and professional skills.
My aspirations lay in progressing my engineering career through pursuit of specialist training and research at PhD level in the field of Ultrasonic Engineering and achieve Chartership status with the IMechE. FUSE CDT is the first academic ultrasonic engineering programme worldwide which would provide me with an exciting opportunity to become a subject-matter expert in this field. Additionally, the programme linked their training to the Monitored Professional Development Scheme (MPDS), which will ensure opportunities, monitoring and feedback are available to allow me to apply for CEng. This is a gold standard of excellence across industry and academia and highly valued by partners and clients.
My project will explore the alignment of emerging technology with automated inspection of challenging industrial inspections. This can take the form of components with complex geometries or restricted access to the entire component surface. Importantly, in many such scenarios, radiography is used for inspection, necessitating clearance of personnel from the vicinity and radiation exposure risks to the inspector. Hence, a move towards ultrasonic techniques would be advantageous from a health and safety perspective. However, such inspection conditions are challenging for standard ultrasonic inspection techniques and the Kelpie sensor approach offers a potential solution for such industrial inspection needs.
My project will consider customised Kelpie array configurations for a number of appropriate industrial components. Simulation will be used to investigate a range of array designs, with close collaboration with Novosound required to align with manufacturing constraints. Extensive characterisation and evaluation of the arrays will be undertaken at the University of Strathclyde to ensure the sensor performance is applicable for NDE inspection applications.
A key objective of my project will be to integrate the Kelpie designs into an automated inspection system capable of operating in areas of restricted access. This task will develop bespoke mechanical systems to manipulate the sensor to ensure full area coverage and subsequently acquire high fidelity ultrasonic data from the component. The final stage will be to investigate post-processing algorithms to produce 3D images of the component under inspection, which will fuse data from the scanning and ultrasonic systems.
Primary Supervisor (University of Strathclyde): Prof. Tony Gachagan
Secondary Supervisor (University of Glasgow): Dr Koko Lam
Project External Partner: Novosound
AI-driven Design of Analog Integrated Circuits for Ultrasound Applications
My name is Yijia Hao. I graduated from the Glasgow UESTC in 2021 with B.Eng. (Hons) in Electronics and Electrical Engineering with Information Engineering. My research interests include deep learning, attention mechanism and application of machine learning in the field of ultrasound.
I am working on ultrasonic circuit design and AI-empowered ultrasonic circuit design tools. The analog front end is essential in most ultrasonic hardware systems, which is widely used in communications, industrial inspection, medical diagnosis, robotically enhanced sensing, surgical tools, etc. Although the analog circuit area is usually less than 20% in a System on Chip (SoC), its required design efforts can be more than 80%. The productivity gap between analog and digital circuits keeps widening. While the design automation level for digital circuits has been increasing steadily over the years, the design automation level for analog circuits remains very low. The analog IC design methodology remains almost unchanged over the past four decades: it is still a slow experience and trial-and-error-driven manual process. Recently, developing novel AI techniques to automate the design of analog ICs starts to attract attention. The AI-driven design lab at the University of Glasgow is a pioneer in AI-driven analog IC design. Being the first few to introduce AI techniques to analog IC design (2008), the AI-driven analog IC sizing method, called ESSAB, was proposed in 2021, which firstly addressed industry-level high-performance analog building blocks (considering the full set of “hard to learn” performances). Through comparison, ESSAB surpasses experienced designers’ design quality in only a few hours. However, process, voltage, and temperature (PVT) variations have not been considered in ESSAB yet. Built upon this, my project aims to:
• Design various analog ICs for ultrasound applications using ESSAB achieving promising results.
• Identify the pros and cons of ESSAB.
• Propose ESSAB-II considering PVT variations obtaining robust designs, which is ready for industry use.
Primary Supervisor (University of Glasgow): Dr. Bo Liu
Secondary Supervisor (University of Strathclyde): Dr. Graeme West
Sensor-Enabled Robotised Plasma Arc Cutting System
My name is Aasim Mohamed. I am a Mechatronics Engineer with a few years of experience in steel fabrication industry, dealing with industrial control systems (ICS), sensors and instrumentation devices. Thanks to that experience, I was able to upscale my career and perceive the complexities within manufacturing and measurement fields. I hold a BSc (Hons) in Mechatronics Engineering from the Future University in the Republic of Sudan, and an MSc in Mechatronics Engineering at De Montfort University in Leicester with distinction. During my master’s, I gained more knowledge and interest in smart systems, robotics, machine vision and flexible automation.
Plasma is a high-energy ionized gas, which is identified as the fourth state of matter. We typically think of three states of matter: solid, liquid and gas. The difference between these states is their relative energy levels. If we take ice as an example, after adding energy to a certain level in the form of heat. The ice melts to form water and then vaporizes into a gas. If you were to add considerably more energy to the steam to heat it up, the steam would break up into several component gases and would become ionized or electrically conductive. The plasma system can produce temperatures approaching 22,000 °C with a velocity that can reach the speed of sound. In comparison, the surface of the sun is about 5,500 °C. Fascinating!
The plasma arc system takes advantage of that to melt and expel material to cut conductive metals. The system can cut at high speed with precision and low cost, which makes it ideal to be used for steel structure applications, shipbuilding and repair, decommissioning process, pipe profiling and weld joints preparation. However, there are some disadvantages. Plasma cutting may require secondary processing to remove the heat-affected material. Also, depending on the job, the plasma machine may require additional setup changes, which could be costly and time-consuming. Metal fabrication has played a vital role in the technological advancement of humankind. There is always a need for improvement in the automation and digitization of manufacturing processes, and demands for moving towards the fourth industrial revolution 4.0.
My research focuses on developing a sensor-enabled robotised plasma arc cutting system. My project aims to optimize the plasma cutting systems for weld joint preparation, repairs, and general cutting applications. The current manual and semi-automated techniques, resulting in higher costs and lower productivity. Ultrasound technology can improve the process by enhancing operational safety, quality control, and production efficiency while digitizing the process. By deploying ultrasound sensors technology to automate parameter adjustments and ensure precise cuts, the process can be optimized in terms of accuracy, cost, productivity, and complex shape cutting.
Primary Supervisor (University of Strathclyde): Prof. Charles MacLeod
Secondary Supervisor (University of Glasgow): Prof. Patrick Harkness
Project External Partner: Kuka Robotics
Development of Artificial Intelligence tools for therapeutic ultrasound
I’m Agnes. My background is in software development and IT, specifically in marine IT systems. After putting my seafaring days behind me, I graduated from University of Glasgow with a BEng in Electronic and Software Engineering in 2021.
My project will involve looking at what preliminary steps would be required to apply AI/Machine Learning (ML) to the Acoustiic system. It is envisioned that the overall project will provide a series of AI/ML tools, each building on the last, that will improve the performance of the Acoustiic system. There are existing data sets available for some of these tasks, while Acoustiic would work with FUSE to develop necessary datasets for others, and additional datasets to enhance existing.
Planned development tasks, that could make up this project, could include:
• Fusion of MR and US images for next generation system, and use of fused images to improve knowledge of material properties of tissue
• Use of ML to improve focusing of ultrasound through aberrating tissue to ensure healthy tissue is not damaged, with the fastest destruction of tumours
• Use of ML to calculate thermal effects of therapeutic pulses for real-time therapy monitoring
• Use of ML for motion tracking of target regions, to remove breathing induced errors in therapy
The requirements for safe, effective USgFUS are extensive and there are many opportunities for ongoing work in this area.
Primary Supervisor (University of Glasgow): Dr. Kevin Worrall
Secondary Supervisor (University of Strathclyde): Dr. Gaetano di Caterina
Project External Partner: Acoustiic
Personalised pulsed-ultrasound to promote bone fracture healing
Hello, I am Andrea Orthodoxou and I am in the first year of the FUSE CDT program as a postgraduate researcher. I graduated from the University of Glasgow in 2021 with an MEng in Biomedical Engineering. During my time as an undergraduate I have had the chance to expand my knowledge and interest in ultrasonics. It is fascinating how the oldest imaging modality can be used for so many different applications in so many different fields.
My project will make use of Polytech’s established in vitro experimental systems and state of the art laser Doppler vibrometry to investigate the micro-motion and nano-displacements induced by low-intensity, pulsed-ultrasound waves as they interact with osteoblast cells lines and bone tissue phantoms. Acoustic simulation (OnScale) will be used to establish the resulting forces imposed on the cells and tissues involved and how these are influenced by the propagation environment (anatomy) and drive parameters employed (frequency, acoustic pressure, etc). With clinical support, the student will use the data gathered to design more effective ultrasound devices, capable of delivering controlled low-intensity, pulsed-ultrasound fields to promote fracture healing across a range of presentations. In addition to the clear prospects offered by successfully advancing the technology, we anticipate potential commercialisation opportunities once we can demonstrate that optimisation of the field to suit the individual presents a significant healing advantage.
Primary Supervisor (University of Glasgow): Dr Helen Mulvana
Secondary Supervisor (University of Strathclyde): Prof. James Windmill
Project External Partner: Polytec
Bio-Inspired Ultra-Compact Acoustic Transmitter - Bubbles as a Means of Weak-Source Amplification
In 2021 I completed my EEE Bachelors at Strathclyde University and immediately after applied for FUSE CDT due to the overlap in concepts with my RADAR based project.
There is a growing need for underwater acoustic devices that are small in size compared to the wavelength that they produce. Current methods for high-power transmissions at lower frequencies rely on the use of large displacements or a large radiating area which is not feasible for a truly compact device. Work has already been done at the Strathclyde Department of Electronic and Electrical Engineering in the Centre for Ultrasonic Engineering in conjunction with Thales Maritime Mission Systems studying the lesser water-boatman male insect (Micronecta scholtzi ) which communicates through efficient generation of underwater sound despite being much smaller than the wavelength of its output. This is achieved by the animal using its air-supply as a secondary amplifier. My project is to replicate this phenomena.
Primary Supervisor (University of Strathclyde): Prof. James Windmill
Secondary Supervisor (University of Glasgow): Dr. Paul Prentice
Project External Partner: Thales
Water Ultrasonic Flow Metering
I studied mechanical engineering during my Bachelor’s. During those 4 years, I learned various mechanical, basic knowledge like solid mechanics, fluid dynamics, and so on. After my graduation, I went to work, first, as project manager, controlling factory’s product quality and process control. Later I worked as a mechanical engineer designing mechanical parts. Then, I started my Master’s study, focusing on non-destructive testing especially in ultrasound technology.
My areas of research interest include non-destructive testing, medical use, and new UT instrument design. There is a growing need to capture multi-parameter data for a new generation of high-performance ultrasonic water metering technology. Physical parameters including pressure, temperature, water quality, and the internal geometry of meters all influence the quality of measurement data. This project will investigate innovations in ultrasonic transducer and meter designs to drive a new generation of ultrasonic water meters, focusing on how multiple physical properties can be captured by a single device, to build a complex picture of the flow measurement environment.
My project aims to engineer ultrasonic transducers (or devices incorporating ultrasonic transducers), able to simultaneously capture multiple high-quality data streams, across pressure, temperature, (accurate) time of flight, and others which can be identified as the project progresses. My objectives are broad, and they can be refined for the benefit of the student. For example, some technical scoping / literature review will be vital in the early stages, and I may study how to optimize transducers to reduce crosstalk or interference with other transducers or structures inside the measurement environment.
Tasks include:
· Design and fabricate new ultrasonic measurement transducer concepts for measurement in water with different physical characteristics (such as pressure, temperature, and quality). One option may be a design based on the flexural ultrasonic transducer.
· Investigate and understand the influence of different environmental parameters, such as pressure and temperature, on the dynamic performance of the ultrasonic transducers. Steps to mitigate undesirable influences will be proposed.
· Characterise and understand the impact of the new transducers on high-performance (for example with respect to accuracy) flow measurement. This objective will require the development and promotion of new signal-processing strategies
depending on the environmental fluid.
· Innovate solutions for capturing the physical properties of a system, through fluctuations in the dynamic performance of the transducers. Advanced materials (phase transforming and metamaterials) can be considered to apply – they might provide unique opportunities to capture other data previously not possible, and it would be interesting if a cost-effective (or practical/realistic) solution for incorporating such materials could be proposed.
Primary Supervisor (University of Glasgow): Dr. Andrew Feeney
Secondary Supervisor (University of Strathclyde): Prof. Tony Gachagan
Project External Partner: Honeywell – Luton, UK
Metamaterials applied to biomedical ultrasound
I am a MEng Biomedical Engineering graduate from the University of Glasgow with a keen interest in the use of ultrasonic techniques within a medical field, particularly within surgical procedures and investigating how tissues respond to ultrasound. During my Masters project I also gained experience in biomaterial formulation and mechanical testing. I am therefore eager to study the materials involved in ultrasonic surgery and how these can be optimised. I am hoping to find a project that combines these interests with the aim of combating challenges such as the improvement of post-operative outcomes for patients and providing ease of use for surgeons through potential ‘smart’ materials.
Metamaterials are those which have been engineered to have properties not found in natural materials. They are made up of repeated structures called unit cells that are constructed from standard materials. It is the structure that determines the bulk material properties, as opposed to the innate properties of the material used. Acoustic metamaterials in particular have the potential to enhance sound production and manipulation. This is because the unit cells can replicate the presence of atoms or molecules, having a size close to that of the wavelength of interest.
For biomedical applications of ultrasound, the benefit of acoustic metamaterials is that they have the capability for small scale design while still operating over a wide range of frequencies. This enables the production of miniaturised transducers, such as those used in catheter ultrasound probes, that do not experience a compromise in performance. Acoustic metamaterials could be employed in many components inside an ultrasound transducer, including the backing and matching layers, as well as the active material that drives the acoustic output. The properties of these components could also be enhanced, beyond what is achievable with standard acoustic materials.
There is an abundance of theoretical studies into metamaterials in the literature, but to be able to incorporate these materials into ultrasound transducers, more practical studies need to be carried out.
The objectives of my project are:
- To explore the existing knowledge on both acoustic metamaterials and biomedical transducer design, with a view of fusing this knowledge.
- To understand the design requirements and constraints involved in creating biomedical ultrasonic transducers and identify key themes for improvement.
- To design and create metamaterials that could potentially improve biomedical ultrasound performance and size.
Primary Supervisor (University of Strathclyde): Prof. James Windmill
Secondary Supervisor (University of Glasgow): Dr. Andrew Feeney
Project External Partner: GE – Nice, France
Automated Ultrasound Data Processing for Defect Detection and Characterization Through Machine Learning
My name is Vedran and I am a PhD student at the FUSE CDT. In 2019 I completed my master’s degree in Power and Process Engineering at the University of Zagreb. To reflect my passion for music and sound, I also pursued a degree in Audio Engineering and Production. Following my work as an engineer, I took the opportunity to move to Scotland and start my academic journey in a new and challenging field.
My project will explore the implementation of automated Phased Array Ultrasonic Testing (PAUT) data interpretation for CFRPs through two approaches:
• developing a low latency Deep Neural Network (DNN) to analyse the A-scan data on the fly, while the scan is being performed, for geometrical feature recognition, automated gating of time series, and defect detection.
• developing a Multitask Network (MN) for image analysis, detection of geometrical features/defects on each B-scan, D-scan, and C-scan projection, and cross-validation of findings at the combination stage.
The real-time DNN applied to the A-scan data will serve to provide warnings for defects flagged during the inspection while the MN, with a potentially improved learning through different related tasks empowered by the multi-view analysis of the data, should be able to detect the defects with higher confidence.
When I am not busy with my studies, you will probably find me exploring Scotland, enjoying music, or grabbing a cup of coffee near the botanical gardens with my family.
Primary Supervisor (University of Strathclyde): Dr. Ehsan Mohseni
Secondary Supervisor (University of Glasgow): Prof. Sandy Cochran
Project External Partner: Spirit Aerosystems – Prestwick, UK