Is there a way of automatically drawing bounding boxes around interested objects?

Question

Given thousands of images, where some of the images contain target objects and others do not, is there an easy way of drawing bounding boxes on these target objects rather than relying on manual annotation? Wouldn't drawing 4 orientations of an object and their respective bounding boxes and randomly inserting them into the images be a viable option?

It becomes painful to manually annotate thousands of images by yourself.

Answer 1

How many classes do you want to annotate it?

We can get the bounding boxes and class names coarsely by using the pre-trained models such as YOLO-V3 darknet, SSD and others.

Then, we can load those annotations using the tool labelImg and manually correct it. It reduces the lot of work and it is called the Human-AI labeling technique.