Mask R-CNN is an example of the image segmentation method based on the depth study, the object can be detected and the target pixel level segmentation.
This hands-on course will teach you to use VIA image annotation tool to create their own data sets and use the Mask R-CNN train their own data set, so that it can carry out its own image segmentation application.
Course Link: https://edu.51cto.com/course/18598.html
The course has three Projects practice:
(1) balloon segmentation Example: in the image to make the detection and segmentation balloon
(2) pothole (single object class) Example segmentation: vehicle driving road scene detection and segmentation pit
(3) roadscene (multi-class object) Example segmentation: vehicle driving road scene pit, vehicles, and detects lane dividing lines,
This course uses Keras version of Mask R-CNN, made presentations on the project Ubuntu system.
This course provides project data sets and python files.
The following is the test results roadscene image segmentation example using Mask R-CNN:
FIG using the Mask R-CNN pothole test results of a single instance of the class object image segmentation:
FIG using the Mask R-CNN of the test results - roadscene class instance object image segmentation: