Finally some connection between AVIF, AI and my position within NVIDIA.
Just published is our blog on using deep learning techniques to help in the Serengeti. Of course this can port over to any national park and wildlife sanctuary in the world.
Jeff Clune heads Evolving AI, out of University of Wyoming and their deep learning-based system can identify species so much quicker than ecologists. The tool will therefore make it so much easier for ecologists to track animals in the wild. As the post states “Camera traps automatically take pictures of passing animals when triggered by heat and motion which produce millions of images”. Clune et al’s research paper shows their system can recognize 48 species in hours vs. thousands of people manually labelling each 6-month batch of images; “saving approximately ∼8.2 years (at 40 hours per week) of human labeling effort (i.e. over 17,000 hours) on a 3.2- million-image dataset.”
The system uses NVIDIA’s DGX-1 AI supercomputer and a convolutional neural network to recognize a variety of animals, with over 92% accuracy.
Lets get the word out to amazing organisations & wildlife conservationists such as Paula Kahumbu.
