Project Description
In this project, a team of students will develop a distributed, machine learning software system for automatically classifying wild animal activity from images captured by multiple, motion-triggered camera traps deployed across 9 square miles of UCSB Sedgwick Reserve. The team will work with computer science, ecology, and sustainability researchers to design tools that inform, automate, and advance multiple disciplines. The system will process, label, and automatically analyze images from multiple locations and investigate model training and classification that spans on-site (low latency) resource constrained systems at the “edge” of the network and resource rich systems in “the cloud”. The images will be processed via computer vision and image processing toolkits (e.g. for object recognition and optical character recognition (e.g. to extract air temperature)). The team will develop (and compare/contrast existing) machine learning models and classifiers that identify animals, count animals, and analyze ecological features and interesting animal behaviors and characteristics. Software development will employ the latest and best software engineering practices (e.g. CI-CD) and the system will include user interfaces, services, and packaging that enables sharing of data, code, and results (e.g. with researchers across disciplines). The result will be a working/tested system that can be deployed and used at other UCNRS sites and beyond. The early work on Where’s the Bear (led by an undergraduate in 2017) is described in this publication and this presentation.
Team Members
- Kaiwen Tang
- Angel Gutierrez
- Sangita Kunapuli
- Jesus Oviedo
Professor and Mentors
- Prof. Chandra Krintz
- Prof. Rich Wolski
Meeting Time
- Research group meeting
- Time and Location: Tuesdays 9AM - noon, TBD
- ERSP group meeting
- Mondays 5PM - 6PM, Wednesdays 3:15PM - 4PM
- ERSP meeting with central mentors (zoom for the first two weeks, in person after that)
- Chinmay: Wednesdays 6PM - 6:30PM
- Diba: TBD
Links to Proposals and Presentation
- Proposal (first draft): link
- Proposal (after peer review): link
- Final proposal (after instructor feedback): link
- Final presentation: link