Object detection - location on image

I’ve got my object detection working for the most part. Now I’d like to extend it a bit to determine where the object is located on the image. For example is the object located on the driveway, street, yard, etc… I’m struggling to figure out a way of doing this.

The only way that comes to mind is to label each pixel in a structure of some sort and use the coordinates of the object to determine where it is. Labeling each pixel will take a long time. Is there another approach? Or an easy way to label each pixel?


Maybe you can try semantic segmentation.
It can cluster an image to the different classes which should meet your requirement.

Below is an example with SegNet for your reference:


This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.