Eliminating false positives during corner finding by merging similar segments
Conference Paper
Overview
Identity
Additional Document Info
View All
Overview
abstract
We present a new corner finding algorithm based on merging like stroke segmentations together in order to eliminate false positive corners. We compare our system to two benchmark corner finders with substantial improvements in both polyline and complex fits. Sketch recognition is an emerging field that utilizes pen-based interaction with computers. Handwriting recognition software in the modern operating systems allows users to write naturally, and applications have been created to recognize sketches in domains such as UML diagrams (Hammond & Davis 2002) and family trees (Alvarado & Davis 2004). In an attempt to perform free-sketch recognition, which allows users to draw as they would naturally without training or being trained by the system, certain geometric sketch recognition systems require a shape to be defined by a set of primitives (Hammond & Davis 2007). Individual strokes can be classified as primitive shapes using Paulson (Paulson & Hammond 2008). However, we would like to allow users to draw multiple primitives in a continuous stroke as they would naturally. Corner detection allows a user to draw in their own style while still allowing a user's drawn shapes to benefit from sketch recognition systems that rely on primitives. In a corner detection system, algorithms automatically break a user's drawn stroke into primitive lines and arcs. This task can be completed during stroke preprocessing, and the resulting primitives can then be sent to a geometric recognizer for stroke classification. We propose Merge CF, a corner detection algorithm that utilizes the stroke's curvature and the user's drawing pen speed in order to find the corners of a stroke. Merge CF then eliminates false positives by removing similar corners, merging like stroke segments together, and examining stroke segments' direction values. Our corner finder is powerful and improves upon current state-of-the-art techniques using two different accuracy measures. 2008.
