Using “caffemodel” files with ROS

You can find here a little video about a project I am working on using “caffemodel” files (which means already trained neural networks with the framework called Caffe [caffe.berkeleyvision.org]).


All caffemodel files can be found on the “caffe model zoo” [github.com/BVLC/caffe/wiki/Model-Zoo] with articles explaining the structure of the neural networks I used and more details about the datasets used to train them.

This step had one objective : to implement several classifiers based on neural networks through the Robot Operating System [ROS] framework. (more details about ROS here : http://www.ros.org/)

GoogLenet : a neural network (more details here : bvlc_googlenet)
NIN : Network in network (a neural network)[the article]
ROI : Region of interest (The ROI node will be replaced by a node selecting good ROIs instead of just proposing the center of the image)

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.