Below is the list of the undergraduate student projects currently being offered by staff in the Department of Medical Physics and Bioengineering during the academic year 2005/2006. The list will continue to grow until the start of term. For more information about the requirements and assessment of the projects, please click here. The important deadlines involved in the assessment of project work are as follows:
Note that Intercalated B.Sc. students must adhere to the guidelines for Physics 4C00 projects. Supervisors and second supervisors are required to keep a record of marks and comments for each component of the project. Assessment forms can be downloaded by clicking on the following:
| Title: | A high performance all-polymer optical fibre sensor for the detection of ultrasound |
| Description: | There is growing safety-related interest in estimating the degree of tissue heating caused by clinical diagnostic ultrasound imaging. Theoretical predictive models have been developed but are limited by the uncertainties in the values of the input acoustic parameters and ultimately the lack of experimental data for validation purposes. There is therefore a need for techniques that can measure ultrasound field parameters and the resulting temperature changes simultaneously and can be used in vivo. The purpose of this project is to contribute towards this aim by developing a probe comprising a thermally and acoustically sensitive interferometric polymer sensing element mounted on the end of an all-polymer optical fibre - the hypothesis is that by matching the material properties of the sensing element and the fibre in this way, there is the prospect of obtaining extremely well behaved frequency response characteristics. The project will involve setting up a laser based experimental set-up, fabricating the optical fibre sensor head and characterising its acoustic and thermal performance. Finally the system will be used to measure the ultrasound induced temperature rise in a phantom that mimics the ultrasound absorption properties of soft tissue. |
| Student: | Elaine Hickford |
| Supervisors: | Dr. Paul Beard and Dr. Edward Zhang |
| Title: | Development of a backward mode laser photoacoustic probe using a piezoelectric PVDF detector |
| Description: | The photoacoustic effect relies upon the generation of high frequency acoustic waves by the absorption of nanosecond near infrared laser pulses. This phenomenon can be exploited to provide structural and functional information about abnormal tissue morphologies, particularly those characterised by changes in the vasculature and local oxygenation status such as cancerous lesions, wounds, and burns. The aim of the project is to construct a photoacoustic probe that could be placed on the surface of the skin and obtain a depth profile of the underlying vasculature. The probe will comprise an optical fibre to deliver the excitation laser pulses and a piezoelectric transducer to detect the resulting acoustic signals. An important aim will be to design the transducer in such a way that the photoacoustic signal can be detected on the same side of the tissue that the laser pulses are delivered to - the so-called backward mode of operation. The project will entail design and construction of the detector, characterising its acoustic performance, designing a suitable probe configuration and testing in tissue mimicking phantoms and in vivo. |
| Student: | Andreas Dorbach |
| Supervisors: | Dr. Paul Beard and Dr. Jan Laufer |
| Title: | The physics of wet skin |
| Description: | This project will involve a combination of making clinical measurements and studying the physics underlying the measurements. Wearers of incontinence pads often have skin problems (abrasion damage, for example) resulting from having their skin in continual contact with wet materials. This project will involve measuring how effective various skin barrier products are at reducing skin hydration, using the model of saline on volar forearm skin. The work will be based at the UCL Archway campus (by Archway tube station). |
| Student: | Esther Kleerekoper |
| Supervisors: | Dr. Alan Cottenden and Dr. Mandy Fader |
| Title: | An incontinence event logger |
| Description: | We have developed to the stage of proof-of-concept a device for logging the times at which an incontinent person leaks urine. This is important in devising personalised care plans. The device uses an array of thermal sensors on the surface of an incontinence pad to detect the arrival of body temperature urine. The next step is to devise algorithms for combining the outputs from the eight sensors in order to distinguish incontinence events from artefacts due to changes in body posture. The project will involve a combination of experimental work and data analysis and will be based at the UCL Archway campus (by Archway tube station). |
| Student: | Dan Reynolds |
| Supervisors: | Dr. Alan Cottenden and Dr. Mandy Fader |
| Title: | Cycling monitor |
| Description: | Electrical stimulation can make paralysed muscles contract. We use this method on paraplegics, disabled by spinal cord injury, so that can propel recumbent tricycles, for sport or recreation. This should keep them fitter and healthier. They do this exercise at home, not in a hospital. However, some supervision is needed to avoid improper leg motion. Because the feet are strapped to the pedals, beside the pedal rotation, the only other possible movement is lateral motion of the knees. Out-of-plane motion is bad and might damage the knee joints. We need a cheap and simple way to measure this sideways motion while the person is cycling with their tricycle on a trainer (back wheel on rollers). We propose to use a simple ultrasound transceiver that is intended to be connected directly to a microprocessor. The project is to investigate this idea. Three parts can be foreseen: (i) setting up the system so that volunteers can cycle while the ultrasound transmitter is pulsed and the responses observed on an oscilloscope; (ii) experiments with different subjects and different clothing to see whether repeatable, meaningful results can be obtained, and how this should be displayed; (iii) a computer or microprocessor can be programmed to demonstrate the system. |
| Student: | Tommy Gorgy |
| Supervisors: | Prof. Nick Donaldson and Tim Perkins |
| Title: | Investigating and compensating for the effect of hair on imaging the brain with optical tomography |
| Description: | A technique called optical tomography has been used at UCL to generate images of blood volume changes in the brains of newborn babies (click here for more information about these previous studies). However, the presence of hair causes errors in the data collection and therefore also in the final images. This project will be in two parts. First, we will build a test phantom (an object with tissue-like optical properties) which includes hair. We will then acquire optical measurements on this phantom and reconstruct images to find out what the effect of the hair actually is. Second, we will investigate and test a new data processing technique which is intended to compensate for the presence of hair. |
| Student: | Sheetal Patel |
| Supervisors: | Dr. Adam Gibson and David Jennions |
| Title: | Data analysis and image reconstruction in optical imaging |
| Description: | A range of projects are available for students with an interest to develop some skills in computing and/or mathematics. These projects concern new optical imaging techniques for displaying tissue function in the brain and other organs. Please click here for a list of projects. |
| Student: | Ali Moghanchi |
| Supervisors: | Dr. Adam Gibson and Prof. Jem Hebden |
| Two projects are being offered which involve using measurements of transmitted light to characterise the scattering properties of the heel bone (with a view to subsequent development as a possible diagnostic for osteoporosis): | |
| Title: | Optical tomography of the adult human heel |
| Description: | The student will build a simple device in which the heel of an adult foot can be placed, surrounded by a volume of scattering liquid (similar to milk). The device will support a series of optical fibres connected to an specialised imaging instrument built at UCL (click here for more information about the instrument and the imaging technique). When completed, the device will be used to acquire data on a volunteer and images of the internal absorbing and scattering properties will be generated. |
| Student: | Amit Gupta |
| Supervisors: | Prof. Jem Hebden and Dr. Louise Enfield |
| Title: | Characterising the optical properties of the human heel and a tissue-equivalent phantom |
| Description: | The student will obtain measurements of the times-of-flight of optical photons across the heels of volunteers, and use them to derive estimates of the average optical properties of the tissue. Experiments will investigate any correlation between the properties and the ages of the volunteers. To test the accuracy of the technique, the student will also make a realistic solid model heel with known tissue-like optical properties (click here for details about how we make tissue-equivalent phantoms). |
| Student: | Nick Norton |
| Supervisors: | Prof. Jem Hebden and Dr. Adam Gibson |
| Title: | Evaluating the use of optical topography for sentinel lymph node detection |
| Description: | During surgery of the breast on women with cancer, it is common to identify and remove the sentinel lymph node. This node, located under the arm, is part of the lymph system that drains fluid away from the breast. However, cancer cells can break away from a tumour in the breast and spread to other parts of the body via the lymph system. Consequently the sentinel lymph node is more likely than other lymph nodes to contain cancer, and it is often removed with the tumour. Breast surgeons at UCL commonly inject a blue dye into the breast prior to surgery, which drains away via the lymph system.Enough dye remains in the lymph system to enable the lymph node to be easily idenitified during the operation. However, surgeons would also like to be able to know exactly where the node is before they start. This project involves testing whether a new technique known as optical topography could produce images of the distribution of blue dye below the skin before the surgeon makes the first incision. First the student will build a solid tissue-equivelent "phantom" has optical properties similar to real tissues. This will contain regions through which liquids can flow at known rates, and at different depths. An existing optical topography system built at UCL will then be used to generate images of the phantom while blue dye is injected into tubes which flow fluid through the phantom. The optical topography system (click here for details) has an array of optical fibres, which allows measurements to be made of near-infrared light which scatters between different locations on the surface. These measurements can be converted into images using computer programs developed at UCL. The ability to identify a lymph node using this technique will be assessed. This project is particularly suitable for an Intercalated BSc student. |
| Student: | Alice Duncan |
| Supervisors: | Prof. Jem Hebden and Dr. Nick Everdell |
| Title: | Ambulatory recording of nerve conduction |
| Description: | Nerve conduction studies are a well established technique in Clinical Neurophysiology to record the conduction of nerves. They are abnormal when nerves are damaged by local mechanical compression or more general biochemical abnormalities. However, their current usage is only is well established under conditions lasting a few weeks or more and studies are performed at rest. There is a category of abnormality, often occurring in sportsmen, where nerves become painful on use, probably because of local compression by expanded muscles, which causes nerve ischaemia (loss of blood supply). At present, there is no established method to test this. The project will be to a) review the relevant literature, b) undertake studies using conventional nerve conduction equipment in a series of normal volunteers during exercise, such as running on a treadmill or stepping, and if time permits c) design and test new electrodes or equipment to overcome any problems encountered in the initial studies. |
| Student: | Andrew Holgate and Alexandros Koukkoullis |
| Supervisors: | Dr. David Holder and Prof. Nick Donaldson |
| Title: | Measurement of cerebral venous saturation using near infrared spectroscopy (NIRS) |
| Description: | Changes in oxy-haemoglobin (HbO2) and deoxy-haemoglobin (HHb) in the brain can be measured using near-infrared spectroscopy (NIRS) instruments. In actual measurements, the HbO2 and HHb signals are often subject to movement artefacts which may be caused by changes in venous blood. The aim of this project is to attempt to exploit these movement artefacts to calculate the cerebral venous saturation non-invasively. The project involves making physiological experiments on volunteers and performing data analysis using the Matlab software package. |
| Student: | Ashley Horne |
| Supervisors: | Dr. Terence Leung and Dr. Clare Elwell |
| Title: | Understanding patient motion in magnetic resonance imaging |
| Description: | Although magnetic resonance imaging (MRI) can provide excellent spatial resolution, this is often limited by the long scanning time, commonly several minutes. During this period, movements - both voluntary and involuntary, can occur, and as a consequence motion artifacts are produced. This project will investigate the effect of physical motions on MRI scans, with an emphasis on exploring whether motions along different axes produce different degrees of artifact. The primary aim is to find out which axis (or combination of axes) produce the worst artifact (ghosting). The possibility of identifying a potential artifact-causing motion by interpreting the data alone will be explored (i.e. without using any tracking methods during the MRI scan). Finally, an attempt will be made to derive an algorithm to correct the data. |
| Student: | Jason Palman |
| Supervisors: | Prof. Roger Ordidge and David Carmichael |
| Title: | Making MR images sharper and less distorted. |
| Description: | The student will explore some new image signal acquisition strategies that minimise the effect of image distortion whilst retaining good signal-to-noise ratio in magnetic resonance (MR) images. Applications include image registration and fusion and interventional image-guided surgery using MR imaging. |
| Student: | Mahrukh Qureshi |
| Supervisors: | Prof. Roger Ordidge and David Carmichael |
| Title: | Design and construction of a breast phantom with realistic x-ray, optical, and elasticity properties |
| Description: | This is a combined project between the Radiation Physics Group and the Biomedical Optics Research Laboratory. The two groups are collaborating on a joined project aimed at building a combined optical/x-ray mammography system. To test the system, we need a test phantom with realistic optical and x-ray properties, which also behaves realistically under mammographic compression. The student will investigate the optical, x-ray, and elasticity properties of different materials and construct and test a realistic test phantom. |
| Student: | Ben Price |
| Supervisors: | Dr. Gary Royle and Dr. Adam Gibson |
| Title: | X-ray diffraction CT of breast tissue samples |
| Description: | The diffraction profiles of healthy and neoplastic breast tissue are significantly different. This is basically due to the different degree of order of the structures, which is lower in tumour than in healthy tissue. Momentum transfer is a function of both the beam wavelength and the scattering angle selected, and if you choose a proper combination of wavelength and angle you can maximise the difference in signal between the two tissue types. The best way to do that is with a synchrotron source, from which you can obtain monochromatic beams (i.e., containing only one wavelength): you tilt the detector at a certain angle and you change the value of the momentum transfer by changing the beam wavelength When using a polychromatic beam, such as the one from a conventional x-ray tube, you must sum the signal obtained from all the wavelength components (each one corresponding to a different value of the momentum transfer for a fixed angle): the risk is that the peaks in the plot are broadened and you no longer have a value of the momentum transfer for which the difference in signal from the different tissue types is significant. You can narrow the x-ray spectrum by using either an appropriate filter or a crystal which "reflects" a narrower energy band. Diffraction CT can provide quantitative information about the diffraction properties of tissues provided they are combined with appropriate transmission tomograms. The goals of the project are: a) characterization of different tumour types, b) characterization of samples taken at different depths in the same tumour (we expect that the diffraction signal will change, expecting broader peaks in the innermost regions of the tumour), c) comparison of the results obtained with the x-ray tube (using a filter and/or the crystal) with those obtained with synchrotron radiation. |
| Student: | Andreas Koudounas |
| Supervisors: | Dr. Silvia Pani and Dr. Gary Royle |
| Title: | Time resolved diffraction imaging of muscle contraction |
| Description: | The contraction of muscle tissue can be observed during an x-ray diffraction study due to the fact that contraction modifies the lattice structure of myosin, a key component of muscle tissue. A number of researchers have already observed this effect in small bundles of muscle fibres and have produced dynamic images that demonstrate the contraction and extension. The aim of these studies has been to study the dynamics of muscle contraction. The proposed project aims to translate these observations to in-vivo measurements in humans. A full description of this project is available here. |
| Student: | Joel Beider and Mark Michel |
| Supervisors: | Dr. Gary Royle and Prof. Nick Donaldson |
| Title: | Early detection of breast cancer using X-ray diffraction |
| Description: | Recent results have indicated that breast cancer tumours as small as 4 mm could be detectable using diffraction enhanced breast imaging (DEBI) techniques. We are currently building a system at UCL that will take advantage of this technique. As part of the programme of work we need to develop techniques to study in detail the diffractive properties of breast tissue samples and this project will consider one of those techniques - microCT. The project will consist of building a microCT system, imaging different samples and comparing the results to standard histopathological analysis. We have in the laboratory a unique, microfocal mammography X-ray source and several different digital detector systems that could be used to build the system. Software exists to reconstruct the data into images. This project is suitable for one student, or two students working together. |
| Students: | Nirmal Perera, Setareh Chavoushi, and Emily White |
| Supervisors: | Prof. Robert Speller and Emily Cook |
| Title: | Simultaneous measurements of brain haemodynamics over the motor and frontal cortices and systemic changes during rest and functional stimulation. |
| Description: | Near-infrared spectroscopy (NIRS) has been shown to measure changes in oxy-haemoglobin (HbO2) and deoxy-haemoglobin (HHb) in the human brain during functional activation, following various kinds of stimulation including motor, visual and cognitive. It has been previously reported that heart rate increases during functional activation of the motor cortex, and it was recently noted that blood pressure increases significantly during frontal lobe stimulation. The aim of this study will be to further investigate the systemic changes that occur during functional activation of the motor and frontal cortex by measuring heart rate and arterial mean blood pressure during functional studies. |
| Student: | Presheena Devendra |
| Supervisors: | Dr. Ilias Tachtsidis and Dr. Clare Elwell |
| Title: | Classification of microcalcifications in mammograms |
| Description: | Work has already been undertaken in the detection of individual and groups of microcalcifications on mammograms. Simple parameters such as size, roundness, linear arrangement etc can assist in the classification of the mammogram into malignant or benign. However, so far, no information from other structures in the breast, for example presence of a lesion, segmental distribution etc has been implement in order to use other than purely local parameters. This project will extend the classification and description of clusters of microcalcificaion by including other descriptive parameters associated with other detected structures in the mammographic images. |
| Student: | Sophie Wrighton |
| Supervisors: | Prof. Andrew Todd-Pokropek and Dr. Tryphon Lambrou |
| Title: | Design of an object orientated multimedia patient report |
| Description: | This could be either an architecture study, or a (considerably) simplified implementation. The concept is that a 'cast' of objects of various types represent the raw data, for example text, graphs and video need to be lined together into a report. These are unsubstantiated into playable objects associated with a 'score', or the way in which a report can be read. They are associated with objects containing information about layout, for example, where on the screen they are to be placed, colour, etc. [This needs to be linked to an understanding of so called DICOM hanging protocols. The 'score' can be played from beginning to end, or navigated. Thus concepts of navigation 'goto', 'if' etc., need to be incorporated. The concept is based on the draft report of the CEN standardisation committee on multi-media records. Two phases needs to be considered: creation, and replay. A particular problem is that is synchronisation of different objects for example video sequences for which a tool such as SMIL could be employed. |
| Student: | Jon Cleary |
| Supervisors: | Prof. Andrew Todd-Pokropek |
| Title: | Parallel computing for the reconstruction of magnetic resonance images |
| Description: | The next few years will see a rapid increase in the number of receiver coils used in magnetic resonance imaging (MRI). Current clinical scanners have difficulty processing this data quickly in order to reconstruct images. This project will look at using methods such as the MATLAB Distributed Computing Toolbox to speed up image computation using multiple computers. The project is suited to a student with an interest in computing and magnetic resonance imaging. A cluster of Linux computers is currently being tested and is expected to be available for use in this project. |
| Student: | Project still available |
| Supervisors: | Dr. David Atkinson |
| Title: | Calibration of freehand 3D ultrasound systems (TWO projects) |
| Description: | Freehand 3D ultrasound systems provide a flexible method for acquiring 3D image data using a conventional 2D ultrasound scanner. In this technique, a spatial position sensor is attached to the ultrasound probe and used to determine the 3D position and orientation of captured image slices with respect to a fixed frame of reference. In order to be able to determine the position of an image from the sensor measurements, calibration is necessary. This is done by scanning a phantom which comprises one or more objects with known geometry in a water bath. In the simplest case, the object can just be a point target, such as a the head of a pin, and there tends to be a trade-off between phantom complexity and calibration time; some of the more complicated phantoms allow calibration to be performed in a matter of minutes with little user-interaction. Two projects are available to investigate alternative approaches to the calibration problem: In one project the aim will be to develop a fully- or semi-automatic method for simultaneous calibration and accuracy determination using a simple phantom (this normally required 2 separate experiments). The second project will focus on a new self-calibrating approach in which the calibration parameters are estimated using an intensity-based image registration algorithm. These projects involve a considerable amount of mathematics and programming using Matlab, so would ideally suit a student who is already comfortable in these areas, or a student who is interested in gaining more experience in these areas. The first project will also involve some practical experimentation using one of the 3 research ultrasound scanners owned by the Centre for Medical Image Computing. |
| Students: | Project still available |
| Supervisors: | Dr. Dean Barratt |
| Title: | Using infrared light to investigation the absorption properties of nonwoven felts used in medical applications |
| Description: | Absorbent nonwoven felts are used in a number of medical applications, notably incontinence pads and wound dressings. It is a major objective of the Continence an Skin Technology Group to establish a better understanding of how fluids interact with absorbent materials and build mathematical models which will enable the development of more effective medical products. This project will use a new infrared device to investigate the absorption properties of nonwoven felts by mapping the distribution of fluid in them under a number of equilibrium (e.g. retention under gravity) and dynamic (e.g. horizontal and vertical wicking) experimental configurations. Data from the new device will be compared with both experimental data using other techniques and predictions based on existing mathematical models. The project will be primarily experimental and based at the UCL Archway campus (by Archway tube station). |
| Student: | Project still available |
| Supervisors: | Dr. Alan Cottenden and Dr. Mark Landeryou |
| Title: | The spatial arrangement of fibres within nonwoven felts |
| Description: | Absorbent nonwoven felts are used in a number of medical applications, notably incontinence pads and wound dressings. This project aims to improve our understanding of absorption within fibrous materials by characterising the spatial arrangement of fibres within a nonwoven fabric. The project will result in a three-dimensional virtual reconstruction of fibres within the chosen nonwoven felt. The project will involve some sample preparation, microscopy, image analysis, computer modelling and visualisation. Previously we have had some success in characterising fibre arrangements using cross-sectional slices, and this work will be extended to three-dimensional structures using stereological methods and, if appropriate, other techniques. We have recently purchased a microscopy and image analysis suite that will be used in this project. The experimental work (supervised by AC) will be based at the UCL Archway campus (by Archway tube station) and the theoretical work in the Engineering building (supervised by ML). |
| Student: | Project still available |
| Supervisors: | Dr. Alan Cottenden and Dr. Mark Landeryou |
| Title: | A study of resolution in photoacoustic imaging using Matlab simulations |
| Description: | Biomedical photoacoustic imaging is an exciting new imaging technique that uses laser generated ultrasonic pulses to image differences in optical absorption within tissue. It has the potential to image to a much finer resolution than diffuse optical techniques such as fluorescence imaging and optical tomography. However, the image resolution is limited by several factors, such as the acoustic inhomogeneity of the tissue, the non-ideal response of the ultrasound detector, and the size of the aperture over which the acoustic signals are detected. A numerical model of the detector response has been developed in Matlab, and will be incorporated into a simple model of photoacoustic generation and propagation, which includes the effects of acoustic inhomogeneties on the acoustic signal. Acoustic time series from this model (simulated measurements) will be used to form photoacoustic images using a conventional photoacoustic imaging algorithm. The aim of this project will be to determine the highest image resolution achievable in various imaging scenarios, by simulating measurements using realistic soft tissue properties and aperture sizes, and a typical detector response. Which factors have the greatest effect on the resolution and why? How can we improve the resolution? |
| Student: | Project still available |
| Supervisors: | Dr. Ben Cox and Dr. Paul Beard |
| Title: | A new apparatus for clinical evaluation of SLARSI patients |
| Description: | Functional Electrical Stimulation (FES) refers to the use of electrical signals to generate muscle contractions in the human body to restore functions lost by paralysis. The Implanted Devices Group has developed a series of implants using electrical stimulation of spinal roots to improve urologic functions and provide a possibility for legs exercise. One form of exercise offered to paraplegic patients is cycling on a recumbent tricycle. The implant receives instruction from the external part (called the control box) via an RF link. The control box can either operate on its own, or an extra box can be connected to it, to allow the clinician to manually select the level of stimulation in real time. It is very usefull to evaluate the performances of the patient, and to measure the effect of stimulation of each individual nerve root. The tricycles are specially designed for FES-cycling. They are fitted with a throttle to alter the level of stimulation, and a shaft-encoder to measure the crank angle. The aim of this project is to develop a new "clinical evaluation box" that will be loosely based on an existing apparatus. It will be connected to and communicate with both the control box and the shaft encoder. The student is expected to build the box and design its internal circuitry, so some understanding of basic electronics and maybe some DIY skills are expected. |
| Student: | Project still available |
| Supervisors: | Prof. Nick Donaldson and Anne Vanhoest |
| Title: | Implant Battery Recharger |
| Description: | We hope to develop an implant to restore urinary continence. The implant will have a re-chargeable battery. In order to minimize the inconvenience to the patient, we propose that re-charging should be done with a large coil that is under their mattress. High-frequency current would be passed through this coil at night and power would be picked up by a small tuned coil in the implant. The project is to develop the under-mattress coil and its driving circuit. This will be a Class E output stage and a circuit to excite it at the correct frequency for the tuned coil. The work will include: design of coils based on information from literature; simulation of the circuits; and building and bench-testing the system. This is an electronics project which will require careful thought about circuits with coils and capacitors and current sources. Most time will be spent in the lab making circuits and testing them. It is clear what sub-tasks must be undertaken. Some prior knowledge of electronics is essential, whether obtained from school, university, or a hobby. |
| Student: | Project still available |
| Supervisors: | Prof. Nick Donaldson and Tim Perkins |
| Title: | Efficient implementation of iterative MRI reconstruction methods |
| Description: | Several popular Magnetic Resonance Imaging (MRI) reconstruction methods (e.g. parallel imaging reconstruction) become very complicated when alternative sampling strategies are used. Specifically, since the reconstruction inversion problem can no longer be broken down into a set of simpler problems, it is necessary to invert a very large matrix. This is not possible with direct methods and iterative reconstruction methods are used. Often, time-consuming interpolation steps are applied repeatedly during the reconstruction and an efficient implementation of such algorithms is essential if they are to be practically applicable. Modern MRI systems are also being equipped with more and more receive channels which further complicates the reconstruction process. Consequently, there is a need for a flexible, fast implementation of these methods. At present we have several interesting reconstruction methods that are running in a Matlab based environment. We would like to have an efficient, stand-alone (written in C/C++) implementation of these methods so that we can start exploring more interesting applications involving large datasets. Ideally this implementation would be able to take advantage of parallel computing (Linux clusters) resources that are being installed at UCL. The ideal student for this project will have good math and computing skills. An interest in MRI and signal processing is an advantage. |
| Student: | Project still available |
| Supervisors: | Dr. Michael Schacht Hansen |
| Title: | X-ray calibration for measurement of breast density |
| Description: | Breast cancer is a leading cause of fatality in women, with approximately 1 in 12 women affected by the disease during their lifetime. Various studies have demonstrated a strong correlation between mammographic breast density and breast cancer risk. In many cases the methods used in these studies to estimate mammographic density have been quite crude and inherently subjective. Promising recent developments, however, have produced techniques which should enable accurate measurement of the volume of glandular tissue in the breast. This project will investigate some of these techniques using phantom images obtained during the project from the digital breast x-ray set at Guy's Hospital. The goal is to use these images together with other measurements and empirical data, to estimate degrading factors such as the anode heel effect and x-ray scatter. This knowledge can then be applied to producing a calibration of the x-ray set and this in turn will enable conversion of intensities in real x-ray mammograms to meaningful quantities such as thickness of breast tissue. |
| Student: | Project still available |
| Supervisors: | Dr. John Hipwell and Prof. David Hawkes (Contact Prof. Hawkes; Dr. Hipwell not available before October 26). |
| Title: | Feasibility study of inverse source modelling or low frequency Electrical Impedance Tomography for imaging action potentials in muscle |
| Description: | When muscles contract, an electrical signal passes along the nerve to the muscle, across the neuromuscular endplate, and thence to muscle fibres. It travels along the fibres at about 5 m/s and this causes a series of chemical changes which cause the muscle to contract. This gives rise to electrical fields, which are recorded with a needle electrode placed in the muscle, which is the basis of the well established technique of EMG (Electromyography). These also combine to produce voltage changes which can be recorded on the surface of the muscle. It is possible to map these by recording with a grid of electrodes. It may be possible to deconvolve these surface signals and track back to their origin within the muscle by the techniques of inverse source modelling or low frequency Electrical Impedance Tomography. If possible, these would allow much more accurate diagnosis of conditions where the nerve to the muscle is damaged, or where the muscle itself is damaged by inflammation (neurogenic or myopathic changes). The project will initially involve reviewing relevant literature on EMG, electrical activity in muscles, and surface mapping of the EMG. The student will then produce analytic or numerical models of the muscle fibre electrical activity and analyse whether the surface signals have sufficient signal to permit reconstruction of the deep sources within the muscle. If time permits, they will then undertake some pilot studies on human volunteers to verify the predictions from modelling. |
| Student: | Project still available |
| Supervisors: | Dr. David Holder |
| Title: | Detection of bone surface direction in ultrasound |
| Description: | An intensity-based registration algorithm has been developed for matching 3D ultrasound (US) images to computed tomography images. The final application for the algorithm is to be used in computer- assisted orthopaedic surgery (hip and/or knee replacement) to allow preoperative information to help guide the surgeon during the operation. The algorithm currently matches by aligning up the position of bone edges in both sets of images. Recent work in image registration has shown the benefit of using directional information during the matching process. We would like to investigate whether information, such as the direction of the normal to the bone surface, can be used to improve the accuracy and robustness of our registration algorithm. The aim of this project is to investigate methods of extracting bone surface direction from ultrasound images. The work will require a review of the literature covering methods of US image segmentation and a general understanding of image segmentation methods. The student will be required to design, implement and test the accuracy of their chosen algorithm. We have excellent data sets available for the student to use, and the facilities to collect more data if required. Depending upon the time available the best method will be integrated with the registration algorithm and accuracy experiments carried out. |
| Student: | Project still available |
| Supervisors: | Dr. Graeme Penney |
| Title: | Methods for assessing the success/failure of registration algorithms |
| Description: | Automated registration algorithms typically employ an iterative method to optimise the value of a similarity measure from a starting position to a final (optimised) position. Unless the algorithm has become trapped in a local minimum the final position should be the "true" registration position. However, the optimisation is iterative and so the algorithms can become trapped in local minima if the starting position is too far away from registration. Therefore, for clinical use, it is imperative that a there is a reliable system to detect misregistrations. There is very little currently in the literature addressing this problem. The aims of this project is to initially carry out a literature survey on current methods, and then to investigate alternative strategies. Possible lines of investigation are: 1) The use of multiple starting positions and then investigating the spread of results. 2) Final values of similarity measures. 3) Investigating possible use of the first differential of the similarity measure during the registration process. 4) The use of more than one similarity measure. |
| Student: | Project still available |
| Supervisors: | Dr. Graeme Penney |
| Title: | Combined diffraction-transmission breast imaging |
| Description: | Mammography is generally done using x-rays to produce an image of the breast under compression. A new imaging system is being developed for mammography in which an x-ray diffraction image is also taken of the breast at the same time as the conventional transmission image. The reasons for the two images is that the transmission image gives very good spatial resolution but fairly poor contrast whereas the diffraction image gives good contrast but at lower spatial resolution. Combining these two images into a single high contrast, high resolution image would maximise the diagnostic information available. This project largely involves manipulating images on a PC using various methods to combine the two different bits of information to optimise diagnosis. |
| Student: | Project still available |
| Supervisors: | Dr. Gary Royle and Prof. Robert Speller |
| Title: | Hair diffraction as an early indicator of cancer |
| Description: | It has been suggested recently that the X-ray diffraction exhibited by hair can be used to detect cancer early. If true this would lead to a very sensitive, screening procedure that could be carried out with little inconvenience to the patient. The results that have been reported so far have all been taken at synchrotron radiation facilities and have shown that the mounting of the hair sample is critical. This project is to investigate whether or not energy dispersive X-ray diffraction (ERDX) could be used to carry out these measurements using a conventional X-ray source. The experiments will be carried out in the UCL radiation laboratories using HpGe or CZT detectors. This project is suitable for one student, or two students working together. |
| Student: | Project still available |
| Supervisors: | Prof. Robert Speller and Dr. Gary Royle |
| Title: | Identifying change in repeated contrast-enhanced MR breast images: Quantification of normal change |
| Description: |
The major goals of breast cancer diagnosis are early detection of malignancy and its differentiation from other breast disease. Traditionally x-ray mammography has been used for this purpose due to its high resolution and sensitivity in fatty tissue and low cost. Its disadvantages include low sensitivity in dense glandular breast tissue and poor signal to noise ratio. An alternative is contrast-enhanced MR mammography. This has lower resolution than x-ray, but provides 3D images and can be applied to dense tissue, post-operative scarred breasts and breasts with silicone implants. The lack of radiation makes it applicable to young pre-menopausal women. We collaborate with the Royal Marsden Hospital, where many MR breast images have been obtained. These are scans of women who are known to be in a genetically high risk of cancer group. A single session involves taking a set of images before and after injection of contrast agent. Between the acquisition of each image the patient may have moved. We have developed and validated a registration algorithm which can be used to realign images before further evaluation. Furthermore, we have many images where the same patient was scanned one year apart and some images taken at two stages of the menstrual cycle. It is important to be able to align images with the corresponding images of previous visits for the purpose of follow up and identification of disease progression. The registration task and the output required are different to the alignment of a volume sequence acquired during a single examination. We still require point by point correspondence of tissue over time but the purpose is to detect changes that cannot be explained by patient repositioning and normal variation in anatomy. In this project we aim to identify a suitable registration method for this task and to compare the changes observed during the menstrual cycle and during one year. The project may entail the following:
|
| Student: | Project still available |
| Supervisors: | Dr. Christine Tanner |
| Titles: |
|
| Description: | A variety of projects are being offered by the Structure & Motion Laboratory at the Royal Veterinary College (RVC). For more details, please download a short description (a WORD document) by clicking here. These projects are based at the RVC, and the first supervisor will be an RVC staff member. However, for any Medical Physics student wishing to take one of these projects, it is necessary that a second supervisor based in the UCL Medical Physics department must be appointed. Any students potentially interested in these projects should first contact the Medical Physics Project Coordinator (Prof. Jem Hebden). |
| Student: | Many projects still available |
| Supervisors: | Various |
| Title: | Title |
| Description: | Description. |
| Student: | Project still available |
| Supervisors: | Supervisor |