This hybrid multiscale deep learning employs a semantic segmentation platform to accurately detect and recognise partially occluded objects. It can be configured for different application needs and is able to perform outdoors and indoors.

 

This robust solution can be adapted for industry applications facing manpower constraints.