Learning Rotational Features for Filament Detection
State-of-the-art approaches for detecting filament-like structures in noisy images rely on filters optimized for sig- nals of a particular shape, such as an ideal edge or ridge. While these approaches are optimal when the image con- forms to these ideal shapes, their performance quickly de- grades on many types of real data where the image deviates from the ideal model, and when noise processes violate a Gaussian assumption. In this paper, we show that by learning rotational fea- tures, we can outperform state-of-the-art filament detection techniques on many different kinds of imagery. More specif- ically, we demonstrate superior performance for the detec- tion of blood vessel in retinal scans, neurons in brightfield microscopy imagery, and streets in satellite imagery.