Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
ML-based multi-subject automated computer vision for chronic in vivo animal trials
In this project, previous work to develop an algorithm that extracts postural data from videos and relates them to acquired data will be expanded upon and evaluated using previously obtained video data.
We are using chronic sheep studies to gather etiological information on Hydrocephalus to be used to steer the design of a novel active cerebral shunt. In these studies, we want to use a trained pose estimation algorithm to detect poses of multiple sheep to gain further insight into acquired data. Machine learning is quickly becoming ubiquitous in academic ventures. Moreover, pose estimation is playing an increasingly large role in chronic studies due to its ability to temporally correlate quantitative and video data. In this project, previous work to develop an algorithm that extracts postural data from videos and relates them to acquired data will be expanded upon and evaluated using previously acquired video data.
We are using chronic sheep studies to gather etiological information on Hydrocephalus to be used to steer the design of a novel active cerebral shunt. In these studies, we want to use a trained pose estimation algorithm to detect poses of multiple sheep to gain further insight into acquired data. Machine learning is quickly becoming ubiquitous in academic ventures. Moreover, pose estimation is playing an increasingly large role in chronic studies due to its ability to temporally correlate quantitative and video data. In this project, previous work to develop an algorithm that extracts postural data from videos and relates them to acquired data will be expanded upon and evaluated using previously acquired video data.
1. Perform a literature review on ML-based computer vision for multiple subjects
2. Identify a desired tool/algorithm compatible with multi-subject pose estimation
3. Customize a dual-camera setup for use with algorithm.
4. Show that the tool can be used to automatically estimate pose for multiple subjects using a dual-camera setup, and that the results can be correlated to acquired pressure data
5. Identify and discuss limitations of algorithm
1. Perform a literature review on ML-based computer vision for multiple subjects 2. Identify a desired tool/algorithm compatible with multi-subject pose estimation 3. Customize a dual-camera setup for use with algorithm. 4. Show that the tool can be used to automatically estimate pose for multiple subjects using a dual-camera setup, and that the results can be correlated to acquired pressure data 5. Identify and discuss limitations of algorithm
1. Cursory knowledge in machine learning, computer vision is a plus 2. Able to navigate large datasets 3. Motivated to expand on your knowledge of applied computer vision in a biomedical engineering setting 4. Desire to contribute to an on-going project that may lead to a fundamentally new biomedical product 5. Able to work independently.
Success in product development depends heavily on the competence and skills of teams and individuals. This is why we dedicate our research to create knowledge that enables the value-adding use of new technologies - and to make this knowledge tangible and teachable. Industrial and clinical needs are the driving forces for our interdisciplinary research. Our work is distinguished by a variety of methods, ranging from simulation to validation of real applications. Our research changes the way we develop products, and our expertise changes
Success in product development depends heavily on the competence and skills of teams and individuals. This is why we dedicate our research to create knowledge that enables the value-adding use of new technologies - and to make this knowledge tangible and teachable. Industrial and clinical needs are the driving forces for our interdisciplinary research. Our work is distinguished by a variety of methods, ranging from simulation to validation of real applications. Our research changes the way we develop products, and our expertise changes
Interdisciplinary research in biomedical engineering
Be a member of the Hydroceophalus Project research group
Gain experience in machine learning and computer vision
pd|z Product Development Group Zürich
Anthony Podgorsak
CLA G 17.2
Tannenstrasse 3,8092 Zürich
apodgorsak@ethz.ch
pd|z Product Development Group Zürich Anthony Podgorsak