4 ms·
How far out are we from Object Detection that can identify at different scales? - i.e. angled, facing towards, rotated, skewed, etc... We seem to have good OD
by prodtorok 9y ago
How far out are we from Object Detection that can identify at different scales? - i.e. angled, facing towards, rotated, skewed, etc...
We seem to have good OD that can create horizontal bounding boxes, but these bounding boxes seem to be generic estimates.
Even rectangle detection with these models can't identify the angle or skew of a rectangle in a frame (we get the same generic bounding box)
OpenCV seemed to have models awhile ago that could do this just fine.
- yorwba 9y agoIf you need an angled bounding box, you could probably modify any of the current approaches in that regard. You could also add a post-processing step where you take the predicted bounding boxes, rotate them in all possible directions, and predict the most likely one. But if you're dealing with known geometric shapes, like e.g. rectangles, you'll get better results if you use "classic" detection algorithms that are already mathematically optimal. For example, I once had to count the number of atoms in an electron-microscope image. I simply ran a circle detector with very sensitive settings, then culling overlapping circles with lower "circleness" score. That missed a few atoms that were stacked on top of others, but still got more accurate results than the previous method, which apparently involved a poor grad student ticking them off on paper.
- alexcnwy 9y agoYou can re-pose it as a regression problem to detect the coordinates of the 4 corners in the case of a single object in the image.
- nshm 9y agoDoes not work since requires much more resources. Also, small objects in high resolution images are painfully slow to detect. Most models are trained for 700x700 at max, 4096x4096 many times slower.