AI Solution for the Confirmation of Joint Bleeds in Patients with Hemophilia using Point of Care Ultrasound - ON-217
Preferred Disciplines: Computer Science (Master’s or PhD level)
Company: AceAge Inc
Project Length: 4-6 months
Desired start date: April/May 2019
Location: Hamilton, ON
No. of Positions: 1
About the Company:
AceAge Inc. is a healthcare technology company, creating intuitive products to ease the aging process and improve health outcomes.
A Feasibility analysis to use a AI solution for the detection of an acute bleed in patients with hemophilia using point of care Ultrasound.
The purpose of this proof of concept study is to develop and test a CNN deep learning algorithm for POC- US scans of ankle joints in patients with hemophilia presenting with suspected joint bleeds. We will start in a small well-define homogenous patient population to answer the question effusion (=blood) vs no-effusion (no-blood) for POC. Second study escalating to more complex heterogenous patient population for POC2.
- Collect and validate the labels of a small set of POC-US scans of ankles for both cases (hemophilia patients with suspected joint bleed) and differential diagnosis controls (non-active joints of patients with JIA);
- Develop a preliminary end toend ML pipeline capable of ingesting, processing, and classifying joints by detection of effusion/ presence of blood;
- Test whether or not the inclusion of presenting clinical features improves the ML classification model;
- Investigate the possibilities of novel hardware to enhance the digital capture of the joint and provide more useable information for the ML algorithm;
- Recommend next steps for the further development of the resulting ML pipeline in order to improve accuracy and validity.
Development of an end to end ML pipeline for ultrasound images
Expertise and Skills Needed:
Machine learning (preferable in imaging), augmented and artificial intelligence.
For more info or to apply to this applied research position, please