Fast Line and Circle Detection Using Inverted Gradient Hash Maps

Abstract

This paper presents fast algorithms for line and circle detection based on inverted gradient hash maps (IGHM). Inverted indices are a common technique for storing a map from content of a dataset to its locations in the dataset. Hash maps are typically used to implement associative arrays and reduce search times in large datasets. In this paper, a hash map is used to store an inverted index of image gradient magnitudes and orientations. Algorithms for detecting lines, and circles using IGHMs are presented and shown to be competitive against existing approaches.