Portable Detection and Characterization of the Synovium and Effusion In Musculoskeletal Ultrasound- ON-489Project type: Research
Desired discipline(s): Engineering - computer / electrical, Engineering, Medicine, Life Sciences, Computer science, Mathematical Sciences
Company: 16 Bit Inc.
Project Length: 6 months to 1 year
Preferred start date: 07/05/2021
Language requirement: English
Location(s): Toronto, ON, Canada; Canada; Canada
No. of positions: 1
Desired education level: CollegeUndergraduate/BachelorMaster'sPhDPostdoctoral fellow
About the company:
16 Bit is a Toronto-based company founded by two radiology physicians with backgrounds in computer science and engineering. With a unique combination of medical and technical expertise, 16 Bit focuses on solving the most impactful clinical problems facing medicine today.
Describe the project.:
Rheumatoid arthritis (RA) is an autoimmune disease often characterized by the inflammation of the synovium and effusion of joints. RA is a chronic condition and may affect quality of life. Detection and characterisation of the synovium and effusion in joints is most often assessed by a trained sonographer or radiologist. Recently ultrasound has been brought into the clinic and used by other trained healthcare professionals such as clinicians and physiotherapists. However, this has introduced more variability to a modality which is already operator dependent. In order to assist in the acquisition of musculoskeletal ultrasound, the current proposal will examine the incorporation of artificial intelligence/ machine learning assisted guidance for scanning to allow for more accessible and standardized assessment. We propose to develop an algorithm to guide the user on how to perform the scan correctly, specifically, where to put the transducer on the joint, what angle, and what pressure to apply. The algorithm will be self-contained and require no intervention from a sonographer. When positioned correctly, the ultrasound device will collect data on the joint and run it through our previously developed image recognition algorithm to characterize the effusion and synovium of the joint. The output will be sent to the clinician for diagnosis and treatment prescription.
1. Develop a registration method to map ultrasound images coming from the transducer. The registration will be performed with various cost functions such as correlation ratio or boundary-based registration.
2. Develop a user interface that would guide the user in the scan. The interface should ideally calculate the best trajectory using a reinforcement learning algorithm and render real-time feedback on the quality of receiving data and guidance of where to move the device, what angle and what pressure to apply.
3. Integrate the solution to the pipeline developed by Prof. Tyrrell and 16-bit for detection and characterization of the synovium and effusion in the knees. The solution should have low latency in guidance feedback and reasonable latency on sending the diagnosis to the clinic.
Candidates should hold a PhD in computer science, biomedical engineering, mathematics, or an associated program, with a track record of excellence and publications in medical image processing and deep learning. The successful applicant will be responsible for developing advanced image processing methods using deep learning techniques for ultrasound imaging for computer-aided diagnosis support and clinically-oriented research. The postdoctoral fellow will also be expected to take a lead on manuscript writing and preparation, presentation of scientific material at national and international meetings, support of grant applications, and mentoring junior lab members. The successful candidate will be located in the MiDATA lab space at the University of Toronto (https://www.tyrrell4innovation.ca/). The postdoctoral fellow will be surpervised by Prof Pascal Tyrrell at the University of Toronto. Pascal is the Director of Data Science and Associate Professor in the Department of Medical Imaging with cross appointments to the Department of Statisitical Sciences and the Institute of Medical Science.