Segmentation

Description

We currently use Mask R-CNN to detect and segment the objects of our images. It was coded using tensorflow and trained using MSCOCO, which means that the classes it uses to segment objects are compatible with GuessWhat?!

The segmentation node outputs a SegmentedImage object which contains the class of the object, a box which delimits the object and a segmentation mask. It is is published on the objects topic.

ROS Installation

Run the install script install.sh

Usage

$ rosrun devine_image_processing segmentation.py __ns:=devine