EdgeDetection2 5.2 Hough Transform In this class, we will learn about Hough Transform. How do we find image boundaries (lines)? ● Hough Transform - Hough Transform is a generic framework for detecting a shape/object. - Edges don't have to be connected. - Lines can be occluded. - The key idea is 'Democratic' Detector. → Each image edge votes for the possible line model. ● Image and parameter space We can see the image as parameter.. 2020. 8. 5. 5.0 Defining Boundaries Today, we will learn about how to define boundaries. What are the object boundaries? This picture would be Human annotated boundaries. By using Edge Detection technique, we can get this kind of edge. By using multi-scale edge detection, we can get this picture. Edge strength does not necessarily correspond to our perception of boundaries. These are pictures we can feel more gap between human ann.. 2020. 8. 5. 이전 1 다음