AI, Software, Smart Automation
Using AI for Semantic Segmentation
Livestock management is a growing market for 3D video, but using AI to detect and monitor animals from depth video is a particularly challenging problem given the lack of labeled data sets and pre-trained AI models. So to do so, one needs to collect data, label the data, train the AI model, and then export the AI model to something one can integrate into OpenCV or OpenVino libraries. Each of these steps is essential to achieving accurate segmentation, and if not given the appropriate time and energy, could spell doom for the entire project. So in this presentation, I share personal experiences with training an AI model from a livestock video and relate what worked and what didn't so you can avoid the mistakes while enjoying my successes in your projects.
Daniel Lau, Professor, University of Kentucky
- It seems you are currently not logged in.
Log in now to confirm your membership level has access to content.
Non-Members can log in for limited access, but logged in members enjoy full access.
If you're not yet a member, click here to join or schedule a meeting with Julia Stevenson today!
*Assets marked for members only are subject to membership level eligibility.
- Platinum
- Gold
- Silver
- Bronze
- Non-Member
