VIAME builds on top of a number of different algorithm and software frameworks, in some case using them as-is, in others having custom implementations with a few modifications.
System Backend - Dawkins, Matthew, Sherill, Linus, et al. "An open-source platform for underwater image and video analytics." 2017 IEEE Winter Conference on Applications of Computer Vision (WACV). IEEE, 2017.
Video Reader - Bellard, Fabrice. "Ffmpeg multimedia system." FFmpeg. [Last accessed: November 2015]. https://www.ffmpeg.org/about.html (2005).
Deep Learning Framework - Paszke, Adam, et al. "Automatic differentiation in pytorch."(2017).
Model Training Harness - Paper forthcoming
Cascade Faster R-CNN Implementation - Chen, Kai, et al. "MMDetection: Open MMLab Detection Toolbox and Benchmark." arXiv preprint arXiv:1906.07155 (2019).
YOLOv3 Implementation - Redmon, Joseph, and Ali Farhadi. "Yolov3: An incremental improvement." arXiv preprint arXiv:1804.02767 (2018).
Rapid Model Generation Implementation - Paper forthcoming
Default Tracker - Sadeghian, Amir, et al. "Tracking the untrackable: Learning to track multiple cues with long-term dependencies." Proceedings of the IEEE International Conference on Computer Vision. 2017.
There have been a number of contributors to VIAME including, but not limited to: (Software Framework and Algorithms) Matt Dawkins, Jon Crall, Linus Sherrill, Matt Leotta, People from the Above References; (Graphical User Interfaces) Matthew Woehlke, Jacob Nesbitt, Brandon Davis, Bryon Lewis, Rusty Blue, Betsy McPhail, Aashish Chaudhary, Kyle Edwards, Matthew Ma; (Detector Model Development) Deborah Hart (Skate), Yuval Boss, Joe Redmon (Arctic Seal, YOLO), Neel Joshi, Marcel Simon (Arctic Seal, TF), and multiple annotators who have generated training data across multiple organizations.