Try run additional lstm on prediction boxes to make it stable from frame to frame.

Real-time and video processing object detection using Tensorflow, OpenCV and Docker.
3.2K
25
Léo Beaucourt
Andrey Nikishaev
·Follow
1 min read·
Apr 14, 2018
--
Try run additional lstm on prediction boxes to make it stable from frame to frame.
--
--
Written by Andrey Nikishaev4.2K Followers
·1.3K Following
Machine Learning and Computer Vision Researcher. Founder LearnML.Today
Responses (1)
Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams