Detecting corner points for the digital images is based on determining significant geometrical locations. Corner points lead and guide for providing significant clues for shape analysis and representation. They actually provide significant features of an object, which can be used in different phases of processing. In shape analysis problems, for example, a shape can be efficiently reformulated in a compact way and with sufficient accuracy if the corners are properly located. This chapter selects seven well referred algorithms from the literature to review, compare, and analyze empirically. It provides an overview of these selected algorithms so that users can easily pick an appropriate one for their specific applications and requirements.
TopIntroduction
Corner detection is a computer vision approach which is used to extract certain kinds of features and infer the contents of an image. It is very frequently used for various applications in real life. These applications include object reconstruction, motion detection, image mosaicking, panorama stitching, image registration, video tracking, object recognition and others. “A corner can be defined as the intersection of two edges. A corner can also be defined as a point for which there are two dominant and different edge directions in a local neighborhood of the point. An interest point is a point in an image which has a well-defined position and can be robustly detected. This means that an interest point can be a corner but it can also be, for example, an isolated point of local intensity maximum or minimum, line endings, or a point on a curve where the curvature is locally maximal” (Wikipedia, 2020).
In general, most of the corner detection methods detect interest points practically. As a matter of fact, the term “corner” and “interest point” are used interchangeably (Willis and Sui, 2009). Hence, if only corners are to be detected, it is necessary to do a local analysis of detected interest points, this will determine which of these are real corners. “Examples of edge detection that can be used with post-processing to detect corners are the Kirsch operator and the Frei-Chen masking set” (Wikipedia, 2020; Shapiro and Stockman, 2001).
Corner detection approaches for binary images usually involve segmenting the image into regions and extracting boundaries from those regions that contain them. The techniques for gray level images can be categorized into two classes: (a) Template based and (b) gradient based. The template based technique utilizes correlation between a sub image and a template of a given angle. A corner point is selected by finding the maximum of the correlation output. Gradient based techniques require computing curvature of an edge that passes through a neighborhood in a gray level image.
Corner detection schemes can be broadly divided into two categories based on their applications: gray level (suitable for gray level images); and binary (suitable for binary images). For gray level schemes, for brevity, the reader is referred to (Moravec, 1980; Harris & Stephens 1988; Kitchen & Rosenfeld, 1982; Noble, 1988; 1989; Lindeberg & Li, 1997; Smith & Brady, 1995; Sánchez et al., 2018; Shi & Tomasi, 1994; Förstner & Gülch, 1987; Kitchen & Rosenfeld, 1982; Koenderink & Richards, 1988; Mikolajczyk & Schmid, 2004; Bretzne & Lindeberg, 1998; Lindeberg, 1994, 1998, 2013, 2015; Lowe, 2004; Lindeberg, 1993, 1994, 1998, 2008; Mikolajczyk & Schmid, 2004; Wikipedia, 2020; Lindeberg and Garding 1997). For binary schemes, for brevity, the reader is referred to (Beus & Tiu, 1987; Chetverikov & Szabo, 1999; Freeman & Davis, 1977; Harris & Stephens 1988; Liu & Srinath, 1990; Pritchard et el., 1993; Rosenfeld & Weszka, 1975; Sarfraz et el., 2006, Sarfraz, 2014). Figure 1 provides an overview of various algorithms suitable for gray level images as well as binary images. This figure also reflects various applications related to corner detection algorithms in different times. This chapter is the extension of the article (Sarfraz, 2014) and mainly deals with techniques adopted for binary approach to detect corners.