XClose

Medical Physics and Biomedical Engineering

Home
Menu

MPBE Summer Studentship Opportunities

Summer Studentships offer a stipend for UCL Medical Physics and Biomedical Engineering students to pursue a project developed in collaboration with an academic over a number of weeks in the summer.

Male and female students working with the Da Vinci robot
Each year the Department of Medical Physics and Biomedical Engineering offers a number of summer placement opportunities for undergraduate students. These placements are an excellent way for students to gain project and research experience during the summer months.

This year, we have two types of studentships available; Research Support and also Teaching and Learning support. Applicants are welcome to apply for 2-3 projects from either category. 

These posts are open to all undergraduate students, including those in their final year. Projects will start in early July 2022.

 

Research Support Summer Studentship

These placements are intended to allow undergraduate students to undertake paid work during the summer with academics. The key aims of the studentships are to give students the opportunity to work on a collaborative research project, gain experience developing and writing a research proposal and presenting their research to their peers.

Available projects:

Visualising research: creating a “toolbox” to showcase innovation in X-Ray Imaging

Supervisor: Dr Charlotte Hagen, Email: charlotte.hagen.10@ucl.ac.uk

Project Aim: This project is about creating clear and appealing visuals to help communicate advanced X-Ray Imaging science to a lay audience. The candidate will develop concepts for communicating scientific outcomes effectively and work on the interface between research and graphic design.

Project Summary: UCL’s Advanced X-Ray Imaging group (AXIm) has been developing innovative X-Ray Imaging systems for nearly two decades. Its flagship invention is a device to measure x-ray refraction and scattering alongside the traditional contrast mechanism in X-Ray Imaging (x-ray attenuation). This so-called phase contrast imaging system has been shown to greatly improve image quality in areas including breast cancer imaging, materials science, and tissue engineering. We have now advanced this device into a tomographic modality, enabling the acquisition of phase contrast images in 3D.

While the working principle of the method has been the subject of many scientific publications, we currently do not have material to effectively communicate this key research outcome to a non-expert audience. Engaging non-experts into research is however of crucial importance, first and foremost to foster a healthy relationship between the scientific community and the general public and for maintaining trust in research, but also for inspiring future generations of scientists and attracting talented individuals from traditionally underrepresented groups to the profession.

This project is therefore about developing material for conducting effective public engagement sessions, for example for use in school science events, university open days, or social media. We are envisaging the creation of a “toolbox” of visuals (schematics, graphics, animations) but are open to new ideas for effectively communicating our research. As the project is situated on the interface between research and visual communication, it would suit a candidate with an interest in science communication, outreach, and external engagement.

A background in x-ray imaging, or imaging in general, will be an advantage but is not essential.

Working Mode: Hybrid (a mix of face-to-face and remote working)

In-silico evaluation of optimal geometry for combined selective stimulation and Electrical Impedance Tomography (EIT) of the vagus nerve

Supervisor: Dr Kirill Aristovich, Email: k.aristovich@ucl.ac.uk and Dr Enrico Ravagli Email: e.ravagli@ucl.ac.uk

Project Aim: The aim of this project is to identify the optimal geometry for combined electrical imaging and stimulation of the cervical vagus nerve, for the purpose of improving therapeutic neuromodulation. The use of simulations will allow to assess several different geometrical patterns in a limited amount of time.

Project Summary: Electrical Impedance Tomography (EIT) is an imaging technique which allows reconstruction of conductivity variations in a given volume based on impedance measurements taken at boundary electrodes. The EIT/Neurophysiology lab at UCL MedPhys is world-leading in using EIT for imaging functional activation in neural tissue; recently, the lab pioneered the use of EIT to image the functional activity of fascicles, small bundles of nerve fibres, and the technology of selective stimulation, to steer electrical current to individual fascicles for therapeutic purposes. Both these technologies can help achieve selective neuromodulation of the vagus nerve, a very promising therapeutic target due to its role in innervating multiple organs and the ease of surgical access. However, current studies in animals are performed using two different nerve cuffs for EIT and selective stimulation, leading to experimental problems like the need to expose a larger section of the nerve and the necessity to co-register data collected at physically different locations. More so, the implantation of multiple devices would not be acceptable when translating these methods to clinical use in humans.

This project focuses on investigating by means of computer simulations the ideal approach to merging EIT and selective stimulation electrode arrays into the same implantable cuff. The student will use the software tools previously developed by the lab to investigate the quality of EIT images reconstructed from simulated data using different electrode geometries; for example, EIT can be performed using the existing selective stimulation electrodes, or EIT electrodes can be intercalated with selective stimulation ones. The outcome of this project will have a real impact on the ongoing research activities as it will lead to the adoption of one of several different strategies for future technical development and translational studies.

Working Mode: TBC

Smart Gloves for Surgical Interventions

Supervisor: Prof Adrien Desjardins, Email: a.desjardins@ucl.ac.uk

Project Aim: This project is centred on the development of novel techniques to detect forces exerted by clinicians with their hands during surgery. In this multidisciplinary collaborative project, we will develop and test novel displays for pressure sensors integrated into surgical gloves.

Project Summary: Clinical context; A challenging aspect of surgical training is to acquire knowledge of the optimal range of forces necessary to successfully complete a given procedure. Uncontrolled or excessive tissue interaction forces can lead to damage and intraoperative complications, whilst insufficient forces prevent task completion. Real-time analysis and display of the forces exerted by surgeons on tissues has strong potential to improve surgical outcomes and to enhance surgical education, by providing objective metrics for training and assessment. Obstetric interventions are a particular area of clinical interest, where manipulations can have profound consequences on the newborn baby.

Smart gloves at WEISS; At the UCL WEISS Centre, we have developed smart gloves with integrated sensors. They comprise electronic resistive pressure sensors that are integrated onto surgical gloves, and corresponding electronics for parallel data-acquisition with a custom Labview interface. These developments took place as part of a multidisciplinary collaboration that includes clinical colleagues (Prof Anna David and Prof Dimitrios Siasakos, UCLH) and Mechanical Engineering (Prof Manish Tiwari and PhD student Carmen Salvadores Fernandez). However, display of sensor data remains an unsolved problem that will be addressed here.

Working Mode: TBC

Characterisation of a multi-user ultrasound-guided focused ultrasound system.

Supervisor: Prof Bradley Treeby, Email: b.treeby@ucl.ac.uk and Dr Elly Martin, Email: elly.martin@ucl.ac.uk

Project Aim: To experimentally characterise a state-of-the art ultrasound-guided ultrasound therapy system and establish treatment guidelines for different therapeutic applications, including neuromodulation, tissue ablation, and opening the blood-brain barrier.

Project Summary: As part of a recent EPSRC / faculty call for multi-user equipment, we purchased a state-of-the art ultrasound-guided ultrasound therapy system (installed in March 2022). The system can deliver ultrasound energy, for example, for tissue ablation, drug delivery, and neuromodulation. This is a multi-user system, with a current consortium of 26 interested users across 9 UCL departments. The equipment is hosted and managed by the Department of Medical Physics and Biomedical Engineering, and is accessible to all staff.

The system delivers ultrasound energy using one of two transducers: a low frequency transducer designed for human or large animal applications, and a high frequency transducer designed for small animal applications. In both cases, targeting is performed under ultrasound guidance using an integrated graphical user interface (GUI). The GUI allows setting the ultrasound therapy parameters, including the duty cycle, pulse length, and driving voltage. However, which parameters to use for a given therapy are currently not known.

The goal of this project is to perform a systematic experimental calibration of the system for both transducers and create a user guide. The user guide will be deployed in the form of a wiki accessible to end users.

Working Mode: Face-to-face

Utilisation of Predictive Modelling for the Forecast of Cancer Outcomes

Supervisor: Prof Maria Hawkins (Professor of Clinical Oncology), Lead Supervisor. Email: m.hawkins@ucl.ac.uk

Dr Charles-Antoine Collins-Fekete (UKRI Future Leaders Fellow), Co-supervisor. Email: c.fekete@ucl.ac.uk

Dr Douglas Brand (Clinical Lecturer in Clinical Oncology), Co-supervisor. Email: d.brand@ucl.ac.uk

Project Aim: The student aims will be to gain familiarity in data formats used in medical care delivery, including clinical trials and to experience in applying a variety of predictive modelling techniques to forecast outcomes.

Project Summary: Each year there are around 375,000 new cases of cancer in the UK. The delivery of treatments for cancer care is becoming progressively more complicated, with large quantities of data accumulated in both the diagnostic and treatment phases. Such data spans a wide variety of data types, including clinical, genomics, imaging, radiotherapy and systemic therapy information. In routine clinical care, a clinician can utilise some of this data in decision-making, but the human mind is not capable of simultaneously utilising all data types in evaluating a decision. It is therefore of interest to develop computer-based models which can integrate diverse data types (multi-omic information) in order to predict relevant clinical outcomes.

Our group is currently working on the utilisation of data both from a clinical trial setting and a real world data setting. Clinical trials are characterised by high fidelity, comprehensive data collection, while real world oncology information is often fragmented and incomplete. This poses challenges in terms of model development, which is easier on clinical trial data but which may then perform poorly in a real world setting.

The student will be integrated into our group based at the Medical Physics and Bioengineering department at Malet Place. They will learn about the different data types which are in use in clinical practice. Specifically also gaining in understanding about key aspects of quality assurance which must be undertaken prior to modelling. Experience will be gained in building a data pipeline to pass this data into the models. Depending upon the interests of the student, a variety of predictive modelling methods can be considered. This could include anything from multivariate logistic regression to neural network deep learning. The precise approaches undertaken will also be dependent upon the individual student and their experience in data science.

Working Mode: TBC

The effects of non-invasive electrical stimulation on neurophysiological outcomes

Supervisor: Dr Lynsey Duffell, Email: l.duffel@ucl.ac.uk and Dr Sarah Massey, Email s.massey@ucl.ac.uk

Project Aim: To help continue with the collection and analysis of data in a research study with healthy, able-bodied volunteers, and to work towards publication. The research study aims to determine the effect of 20 minutes of transcutaneous spinal cord stimulation (tSCS), delivered with and without a 10 kHz carrier frequency, on human neurophysiology.

Project Summary: Electrical stimulation delivered over the spinal cord, known as transcutaneous spinal cord stimulation (tSCS), has been shown to promote functional recovery for patients with spinal cord injury. However, this is not always an accepted method of intervention, as the intensity of the stimulation is not always accepted by the patient. The addition of a carrier frequency (in the order of kilohertz) during tSCS may reduce the sensation which is often not accepted by patients. This method may also provide patients with functional recovery, but less research on interventions using a carrier frequency have been carried out. This project investigates the mechanisms by which tSCS interacts with the central nervous system and the effects of tSCS along the path of the muscle of interest for when delivered with and without a carrier frequency.

The outcome measures used in the study are the H-reflex and motor-evoked potentials (MEPs). The H-reflex is a measurable signal from a muscle when a nerve is electrically stimulated. MEPs are signals which are measured from a muscle when the brain is stimulated using transcranial magnetic stimulation (TMS). Using these two measures allows us to provide a better understanding of how the given intervention interacts with spinal and corticospinal pathways. The team have already collected data from 7 participants, and wish to continue this work with further participants and work towards publishing the data.

Working Mode: Hybrid (a mix of face-to-face and remote working)

Medical image processing for OCT

Supervisor: Prof Marinko Sarunic, Email: m.sarunic@ucl.ac.uk

Project Aim: The purpose of this project is to investigate the causes of myopia (short sightedness), and evaluate the effects of novel interventions. The project work is based on image processing of clinical optical coherence tomography (OCT) data.

Project Summary: Short-sightedness, also known as myopia, is rapidly increasing in young people across the UK and around the world. At the early stages of myopia, blurry vision can be readily corrected by wearing glasses. However, as the onset of myopia occurs at a younger age, and as the severity gets worse, the probably of developing vision robbing diseases in later life significantly increases. The aim of this project is to develop imaging technology to understand the physiological changes in the eyes that cause myopia, and to evaluate the effects of proposed treatments.

OCT is a non-invasive diagnostic technique that is commonly used for volumetric (three-dimensional) micrometre-scale imaging the retina, the light sensitive tissue at the back of the eye. Visualization of the choroid, a heavily vascularized tissue that provides metabolic support for the retina, is at the very limit of OCT capabilities.

The quality of OCT images of the retina and choroid can be improved by acquiring and averaging multiple volumes. An introductory step in this project will be to implement volumetric image-registration, compensating for motion that occurs during imaging, and permitting pixel-to-pixel alignment and averaging of the image data. The second part of the project will be to develop software methods to automatically analyse the blood vessel patterns in the choroid. These software tools will contribute to research on the hypothesis that myopic severity is related to a layer of the choroid tissue where the middle-sized vessels are located.

More information here

Working Mode: Hybrid (a mix of face-to-face and remote working)

Developing haptic feedback for Intraosseous Transcutaneous Amputation Prosthesis (ITAP): modelling bone anchoring

Supervisor: Dr Henry Lancashire, Email: h.lancashire@ucl.ac.uk and Dr Catherine Pendegrass, Email: c.pendegrass@ucl.ac.uk

Project Aim: To model the propagation of vibrations through a bone-anchored, transcutaneous prosthesis inserted into bone.

Project Summary: Intraosseous Transcutaneous Amputation Prostheses (ITAPs) provide patients with sensory feedback through osseoperception. However, ITAP users still lack the senses of proprioception and touch through their prosthetic limb. ITAP provides a unique opportunity to provide direct sensory feeback: by sending vibrotactile signals along the device to the patient. This project will investigate the transmission of vibrations from an external prosthesis to the patient through an ITAP device.

The student will further develop an existing model of force transmission through ITAP using finite element modelling. The tasks on the project are as follows:

  1. To develop a model of ITAP in press-fit and cemented fixation within a long bone, and to predict vibrotactile force transmission using this model.
  2. To extend the initial model using a simplified anisotropic bone model, and to predict vibrotactile force transmission using this model.
  3. To develop a clinically relevant ITAP-bone model using clinical scans of bone remodelling to predict vibrotactile force transmission.

Working Mode: Face-to-face

Development of an ergonomic external stimulator system for a bladder control neuroprosthesis for paralysed pet dogs

Supervisor: Dr Henry Lancashire, Email: h.lancashire@ucl.ac.ukDr Sean Doherty, Email: sean.doherty.15@ucl.ac.uk and Prof Nick Donaldson, Email:  n.donaldson@ucl.ac.uk

Project Aim: This project aims to develop an ergonomic system for use as an external controller for a neuroprosthetic to restore Bladder and Bowel Control in paralysed pet dogs.

Project Summary: The student will develop a new ergonomic system for use as an external controller with an existing implanted dog bladder control neuroprosthesis. A present system, used to restore emptying of the bladder and bowel in dogs with spinal cord injury, is challenging for users: it requires two hands to operate, unsuitable when also restraining a dog.

In this project the student will work with a team including industry and veterinary collaborators to redesign an external controller. The project will involve exposure to a scaled-down version of the medical device development process, the student will be expected to interact with the veterinary clinician, research team and industry partner. This project has a short term route to impact for existing users of the veterinary device as well as longer term potential to underpin future research projects

A successful student will have knowledge of 3D Computer Aided Design, 3D printing, electronics assembly, and medical electronics.

Working Mode: Face-to-face

3D printed detector system for directional radiation monitoring

Supervisor: Dr Rob Moss, Email: robert.moss@ucl.ac.uk and Prof Robert Speller, Email: r.speller@ucl.ac.uk

Project Aim: To design, build and test a novel 3D printed detector system.

Project Summary: The ability to localise sources of radiation is essential for a number of applications including in medicine and security. In recent years, UCL has developed a new approach to directional radiation detection called RadiCAL*. The concept utilises a scintillator and a photomultiplier tube. The scintillators is cuboid in shape and is rotated with respect the environment. If a source of radiation is present then the recorded signal fluctuates as the wide face and narrow face are presented to the source. By encoding the scintillator rotation position it is possible to use this information to infer the direction of a radiation source.

In this project we intend to produce a fully functional prototype (TRL 5) detector that can be used to map out and characterise point and distributed sources of radiation. To achieve this we will further develop an existing design concept and manufacture the majority of parts using 3D printing facilities that are available at UCL.

The prototype will be characterised using various sources of radiation. The detector will be an essential tool for ongoing research in the Radiation Physics Group and can be used to validate Monte Carlo (computational) models as well as generate test data for novel data analysis algorithms.

The project will involve training to safely use sources of ionising radiation (radioisotopes in this case) as well as instruction in 3D modelling and 3D printing. There is also a need to become familiar with motion control components and writing short scripts (in Arduino and Matlab) to control movement and synchronise data acquisition.

*see https://ieeexplore.ieee.org/document/7581900 for example.

Working Mode: Face-to-face

Dosimetric characterisation of alanine response in a tissue phantom following clinical radiation treatment

Supervisor: Prof Maria Hawkins, Email: m.hawkins@ucl.ac.uk
Prof Andrew Nisbet, Email: andrew.nisbet@ucl.ac.uk
Dr Reem Ahmad, Email: r.h.ahmad@ucl.ac.uk
Dr Stephen Turnock, Email: s.turnock@ucl.ac.uk

Project Aim: Assess the usability of alanine as a dosimeter within a bespoke phantom comprised of 3D printed material and tissues irradiated with a clinical radiationbeam(proton/X-ray). 

Project Summary: The research group currently conduct experiments with a 3D printed liver phantom that contains special inserts that can hold tissue samples. The radiation response of normal, non-cancerous tissue is of particular interest;however, cancerous tissues are also used. The work looks at both physical dosimetric responses and biological responses. Physical doseresponses are currently characterised through gafchromic film measurements. However, an issue with film measurements is their sensitivity tolight exposure which can alter and affect the reliability of absolute dosemeasurements taken. This project aims to assess the usability of alanine within this setup and determine differences in dosimetric response. With an optimised dosemeasurement protocol,it will be possible to link differences in biological response to radiation doses delivered to the tissue.

Working Mode: Face-to-face

Compact and miniaturized fiber-optic magnetic sensor using magnetic elastomeric composites

Supervisor: Dr Zhi Li, Email: zhi-li@ucl.ac.uk and Dr Sacha Noimark, Email: s.noimark@ucl.ac.uk

Project Aim: The aim of this project is to develop a highly miniaturized fiber-optic magnetic sensor comprising flexible, magnetically deformable elastomeric composites and fibre-optic interferometric sensing technology, to enable wide-ranging biomedical applications such as biological activity monitoring or medical device navigation.

Project Summary: Magnetic field sensors have broad applicability in many biomedical applications such as biological activity monitoring or medical device navigation. However, to achieve high device performance, these magnetic field sensors need to demonstrate a wide range of properties such as high sensitivity, device miniaturisation, good bio- and chemical compatibility, and high resistance to electromagnetic (EM) interference. Fiber-optic magnetic field sensors are well-suited to these applications, compact size, mechanical flexibility, intrinsic material safety and immunity to EM interference. Moreover, by integrating magneto-optical materials with fiber-optics, various types of magnetic sensors can be developed based on magnetostrictive effects, or Faraday effects. The magnetic signal can be readily translated into an optical signal through the well-established fiber-optic interrogation schemes. Nevertheless, current fiber-optic magnetic field sensors can be limited by low sensitivity and reliability due to the poor integration of the magnetic components with the optical fibre or the structural complexity.

In this project we propose the design of novel, miniaturised magnetic sensors using soft magnetic elastomers and fiber-optics. These magnetic elastomers are a class of smart materials consisting of magnetic particles integrated within non-magnetic elastomer networks. Their excellent magnetic and mechanical properties enable fast, untethered, and reversible responses to external magnetic induction. Using fiber-optic interferometric technology, their superior magnetic responses can be exploited for sensing magnetic field with ultra-high sensitivity. These magnetic sensors can be fabricated using dip-coating methods, to directly apply magnetically responsive materials onto the end-faces of single mode optical fibers. This project will focus on optimization of the material development and the sensor design to optimise the magnetic sensing performance (magnetic responsiveness, stability, and reproducibility) to facilitate steps towards development of a benchtop prototype. These highly sensitive, miniature fiber-optic magnetic field sensors are cost-effective, simple in design, immune to EM interference and are well-suited to a wide range of biomedical applications.

Note: this project is a continuation work of one UG project of this year. Current study show exceptional good results which is worthy conducting further and comprehensive investigations.

Working Mode: Face-to-face

 

Teaching and Learning Support Summer Studentship

These placements are intended to allow undergraduate students to undertake paid work this summer with academics to develop teaching and learning materials, content or activities for use in MPBE courses.

Available projects:

Automatic Grip Force Control in a Robotic Hand

Supervisor: Dr Prabhav Nadipi Reddy, Email: p.reddy@ucl.ac.uk and Eve Hatten, Email: e.hatten@ucl.ac.uk

Project Aim: To explore the feasibility of a set of new lab exercises for the Medical Instrumentation 2 module that would require the students to build an automatic grip force control for a robotic hand.

Project Summary: As humans, we can hold an object without paying any visual attention to the task. This is because we can sense when the object starts to slip. If there is a slip, we automatically increase the grip force. Most prosthetic hands cannot detect this kind of slip that involves sensory feedback.

The project aims to develop such an automatic control system to prevent slip in a robotic gripper. The system would involve a sensor that would detect when the gripper is holding an object. Another sensor would detect if this object starts to slip out of the gripper. If slip happens, a control system would change the input to a motor which would increase the grip force till the slip stops.

Once the system is built, we want to explore how the project could be divided into smaller pieces that could be given to students as lab exercises. We want to use these exercises to help students learn and get practical experience in analog electronics, control systems and embedded programming in the context of the Medical Instrumentation 2 module.

Working Mode: Face-to-face

Establishing a local resource of open medical imaging datasets

Supervisor: Dr Jamie McClelland, Email: j.mcclelland@ucl.ac.uk

Project Aim: To create a local collection of cleaned and pre-processed medical imaging datasets from open access sources, that are ready for use as a teaching resource for modules and student’s projects at UCL.

Project Summary: UCL offers a number of taught modules and student projects related to applying image processing techniques to medical imaging applications. The practical exercises, coursework, and projects greatly benefit from the use of real medical imaging datasets, but often require large amounts of data, especially for work that involves modern machine learning (deep learning) methods. Furthermore, it is usually not permitted to use datasets acquired for research for teaching purposes, greatly limiting the data that is available. Fortunately, there are now many sources of large medical imaging datasets which are available from open access online archives (aylward.org - Open-Access Medical Image Repositories). However, since the datasets are from many different studies and institutions, there is great variability, including the type of data (CT, MRI, manual labels, etc), anatomical sites (lungs, brain etc.), image characteristics (file format, size, resolution, etc.), and naming conventions.

This project aims to set-up an in-house pre-processed resource of medical imaging data for teaching and learning purposes. This will involve a number of tasks, including:

  • searching and cataloguing the open datasets that are available,
  • determining which datasets are most suitable for our current and future teaching needs,
  • downloading a number of datasets that are most suited to our needs,
  • applying various data cleaning and pre-processing steps, including:
    detecting missing and anomalous data
    converting data into suitable file formats
    generating label images from contour (point-set) data
    applying consistent naming conventions
  • Arranging, and storing the data so it is ready for use in future teaching purposes.

 

Several academics from CMIC and WEISS have already expressed keen interest in utilising the datasets from this project for exercises/coursework in the modules they teach, including: MPHY0025, MPHY0030, MPHY0041, MPHY0047, and MPHY0026. The availability of the datasets will also be advertised to the wider UCL community through appropriate channels.

Working Mode: Face-to-face

Development of interactive online materials for teaching anatomy & physiology to engineering students

Supervisor: Dr Rebecca Yerworth, Email: r.yerworth@ucl.ac.uk and Dr Lynsey Duffell, Email: l.duffell@ucl.ac.uk

Project Aim: To expand on the interactive online materials that have already been developed on Moodle to teach anatomy & physiology to engineering students, specifically to improve the consistency and increase the content and interactive activities.

Project Summary: Over the past few years, teaching has adjusted to being primarily online due to the pandemic. While most teaching activities can hopefully return to being face-to-face from the next academic year onwards, it has been acknowledged that some topics are better taught by a blended design, where online materials replace traditional lectures, and these are supported by seminars and/or practical sessions.

In the 2021/2022 academic year, we used this approach to teach anatomy & physiology to our biomedical engineering and medical physics students. Teaching anatomy & physiology to our engineering students is important because they work closely with medical devices, used to diagnose or treat a variety of pathologies. Therefore, the students need a broad understanding of how the body works, its structure (anatomy) and its function (physiology). They also need some anatomical vocabulary, to be able to communicate to other engineers and to talk to clinicians.

The course aims to provide a broad overview of Human Anatomy & Physiology, consistent with the syllabus outline produced by IPEM. Topics include:

  • General structure and organisation of the body
  • Anatomical position and nomenclature
  • Surface anatomy
  • Locomotor system – skeleton, head, trunk, limbs, joints, muscles
  • Cardiovascular system – the heart, respiration, blood flow
  • Renal, urinary, gastro-intestinal, reproductive, endocrine systems
  • Neurological system, pathways, nerve conduction, key bio signals

We have already developed interactive materials covering all topics, which the students are encouraged to work through in conjunction with the course textbook. However, the materials were developed by 3-4 different academics, within a short time-frame. Therefore, there is a need to improve the consistency of the materials and to make the content richer by increasing the interactive activities (quizzes, missing words, labelling diagrams) and the links to clinical scenarios.

Working Mode: Online

Machine Learning for Automated Error Detection in Laparoscopic Colorectal Surgery

Supervisor: Dr Evangelos Mazomenos, Email: e.mazomenos@ucl.ac.uk

Project Aim: Motivated by latest developments in learning-based architectures for video analysis that enable the identification and temporal localisation of actions and events in videos, this project will develop AI technology to detect surgical errors in laparoscopic colorectal surgery.

Project Summary: Due to better aesthetic outcomes, shorter hospital stays and faster recovery times, laparoscopic surgery is the preferred method for treatment of colorectal conditions (colon/rectal cancer, inflammatory bowel disease, rectal prolapse, diverticulitis).

However, minimally invasive abdominal access through small incisions in the abdomen results in increased complexity for surgeons as the operation is taking place in a limited workspace and under field-of-view constrains. These challenges combined with other exogenous factors (e.g., fatigue) rise the risk for sub-optimal execution and intra-operative complications, which are not uncommon in colorectal procedures and are connected with post-operative mortality and morbidity (https://dx.doi.org/10.1055%2Fs-0036-1584501). These may range from minor errors (not optimal use of tool/approach used) to sever complications (excessive bleeding due to manual errors) that require immediate attention and significantly affect the workflow of the procedure.

The proposed project is supported by a dataset of 164 fully annotated laparoscopic colorectal cases obtained through a collaboration between UCL WEISS and the Griffin Institute at Northwick Park. The dataset is annotated in terms of surgical errors type, their temporal location and severity and provides an excellent database for developing novel AI technology for surgical video analysis.

The main objective is to develop supervised spatio-temporal machine learning methodologies to automatically detect surgical errors in videos of laparoscopic colorectal surgery. We will focus on multi-level, self-attention mechanics to extract localised feature representations to train an AI architecture to detect moments of errors in the surgical videos.

Working Mode: Hybrid (a mix of face-to-face and remote working)

 

Eligibility

Open to all undergraduate students in Medical Physics and Biomedical Engineering, including those in their final year and due to graduate this summer.

Renumeration

The studentship will pay London Living Wage for up to 8 weeks (London Living Wage is currently £11.05ph, for a 36.5 hour UCL week, we pay £403 per week), typically starting from the beginning of July. Training will be provided appropriate to the requirements of the project. In addition, up to £200 may be requested for materials and consumable items.

Submission Instructions

All applications should be emailed to mpbesummerstudentships@ucl.ac.uk (subject title: ‘Summer Studentship application - insert your name here’).

There is no application form, instead each applicant is required to submit a cover letter (no longer than one page) specifying their top 2 or 3 projects and addressing how they meet the required attributes, their prior experience in research and/or teaching and why they are interested in their chosen project/s. They must also submit a one-page CV.

DEADLINE: Sunday 24th April 2022, 23:59