Nati Kligler, M.Sc. Thesis Seminar
Tuesday, 29.8.2017, 11:30
We introduce the concept of visibility detection within a point set to new domains. Specifically, we show that a simple representation of an image as a 3D point cloud lets us use visibility detection in classical image processing tasks, improving state-of-the-art results. Given an image, each pixel is represented as a feature point, a viewpoint is set, and points that are visible to the viewpoint are detected. What does it mean for a point to be visible? Although this question has been widely studied within computer graphics, it has never been regarded when the point set consists of feature vectors (rather than a real scene). We show that the answer to this question reveals unique information about the image, enabling us to modify state-of-the-art algorithms and improve their own results. As proof of concept, we demonstrate this idea within three applications: text image binarization, document unshadowing and stippling-style illustration.