This informal web conference series is for those working on computer vision for camera trap data.
The first meeting took place on June 3rd, 2019,
The second meeting took place on July 22nd, 2019.
The goal is to discuss common challenges as well as new ideas, to advance the work in this very niche field.
Meeting One
DATE – MONDAY, JUNE 3rd
TIME – 9 AM – 11 AM (PST)
MEETING RECORDING.
SHARED GOOGLE DOC.
Meeting 1 included teams from Rutgers, San Diego Zoo Global, Google/Wildlife Insights, and many more.
Each team gave a short presentation of their current work.
Common topics included the need for standards and best practices, ethics, and keeping the human in the loop.
We also discussed object detectors versus classifiers, problems with transfer learning, different approaches to handling backgrounds, and the effect of sequences on training/testing splits.
ORDER OF PRESENTATIONS
- Rutgers
- San Diego Zoo Global
- Google Earth Outreach / Wildlife Insights
- Tabak et al.
- Yousif et al.
- Microsoft AI for Earth
- Wild.Me
- Zooniverse
- Snapshot Safari
- Wildlife Insights
- Zoological Society London
- iWildCam 2019
Meeting Two
DATE – MONDAY, JULY 22nd
TIME – 9 AM – 11:00 AM (PST)
Recent Works by Attendees
- Recognition in Terra Incognita. Beery et al. 2018.
- Synthetic Examples Improve Generalization for Rare Classes. Beery et al. 2019.
- Pity the analyst: Designing software for image inspection. Greenberg. 2019.
- Counting Giraffes. Lesk et al. 2019.
- Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning. Norouzzadeh et al. 2018.
- An Animal Detection Pipeline for Identification. Parham et al. 2018.
- Machine learning to classify animal species in camera trap images: Applications in ecology. Tabak et al. 2019.
- Responsible AI for Conservation. Wearn et al. 2019.
- Identifying animal species in camera trap images using deep learning and citizen science. Willi et al. 2018.
- A transient search using combined human and machine classifications. Wright et al. 2017
- Fast human-animal detection from highly cluttered camera-trap images using joint background modeling and deep learning classification. Yousif et al. 2017.
Additional Resources
This list, courtesy of Dan Morris, which covers everything related to camera traps and machine learning.
Coverage of the Fine-Grained Visual Categorization (FGVC) workshop at CVPR 2019, which featured the iWildCam 2019 challenge.
Stay in Touch
If you would like to join this group or attend the next event, please fill out the form below.
Contact Ariel.Hammond@Rutgers.edu if you need help accessing the papers.
Ecological Data Management 