Adding small subimages into larger image

I am working on an object labelling system that works on black/white images. As part of the system, I use a correlation system to find duplicates of a given object elsewhere in the image. The final results of this correlation system is a matrix of the same dimensions of the original image, with single-pixel white dots indicating the centroid of the duplicates.

My problem is that I now want to make a label matrix with all of the detected objects, so that different label values correspond to the different types of objects (all duplicates would have the same label). Is there a fast and efficient way to 'imprint' the objects into a label matrix if I know the coordinates of their centroids?

I have the object saved as a binary image the size of its bounding box, is there any way to avoid having to explicitly define the range of matrix indices to overwrite for every single duplicate? I want to avoid this because there is often cases where two objects have an overlapping bounding box or might even be adjacent to each other. I have attached an image as an example of the type of objects I am working with.