Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Object tracking for improved keypoint detection on a computer vision algorithm for in-vivo ovine trials
We have developed an in-house coded computer vision algorithm to be used in posture estimation and pressure correlation for a chronic ovine in-vivo trial. We currently have instituted a HighRes-Net based architecture for keypoint detection but would like to investigate Object Tracking methodology to potentially improve our detentions.
Hydrocephalus is a disease characterized by an excess accumulation of cerebrospinal fluid (CSF). This accumulated CSF can lead to a slew of clinical sequelae, including impaired gait, memory loss, incontinence, etc. To better understand the precursors to this disease, we have conducted chronic sheep trials with a surveillance system for constant observation. The angular postural data from the algorithm s then linked with simultaneously acquired telemetric pressure data for statistical interpretation and further processing. We would like to investigate object tracking for improved keypoint estimations
Hydrocephalus is a disease characterized by an excess accumulation of cerebrospinal fluid (CSF). This accumulated CSF can lead to a slew of clinical sequelae, including impaired gait, memory loss, incontinence, etc. To better understand the precursors to this disease, we have conducted chronic sheep trials with a surveillance system for constant observation. The angular postural data from the algorithm s then linked with simultaneously acquired telemetric pressure data for statistical interpretation and further processing. We would like to investigate object tracking for improved keypoint estimations
First, you will become acquainted with our pose estimation algorithm (documentation and GitLab repos provided). Then, you will take our development pipeline and implement object tracking into the workflow. Once implemented, the keypoint detection results will be quantitatively compared and evaluated against the current setup. If the results provide an improvement, the system will be transferred into our production flow and used for analysis of real data.
First, you will become acquainted with our pose estimation algorithm (documentation and GitLab repos provided). Then, you will take our development pipeline and implement object tracking into the workflow. Once implemented, the keypoint detection results will be quantitatively compared and evaluated against the current setup. If the results provide an improvement, the system will be transferred into our production flow and used for analysis of real data.
Experience in machine learning methods (Python is a must, with experience in computer vision a large plus). You're motivated to work in a truly multi-disciplinary project with real clinical implications. With your analytical and quantitative mind, you're excited by the prospect of being actively integrated into our project team
The chair of Product Development
and Engineering Design at the
ETH Zurich considers itself a center
for system-oriented product
development and innovation. Our
aspiration consists on the one hand
of the advancement and investigation
of methods and processes of
product development and on the
other hand of the development of
new technical systems. The purpose
of our daily work is to contribute
to the innovative ability and
competitiveness of Switzerland.
The chair of Product Development and Engineering Design at the ETH Zurich considers itself a center for system-oriented product development and innovation. Our aspiration consists on the one hand of the advancement and investigation of methods and processes of product development and on the other hand of the development of new technical systems. The purpose of our daily work is to contribute to the innovative ability and competitiveness of Switzerland.
Interdisciplinary research in biomedical engineering and computer science
Active integration into our project team comprised of surgeons, veterinarians, scientists, and engineers. You will gain experience in applied machine learning and see how a large multi-institutional project works.