Corner points or features determine significant geometrical locations of the digital images. They provide important clues for shape representation and analysis. Corner points represent important features of an object that may be useful at subsequent levels of processing. If the corner points are identified properly, a shape can be represented in an efficient and compact way with sufficient accuracy in many shape analysis problem. This chapter reviews some well referred algorithms in the literature together with empirical study. Users can easily pick one that may prove to be superior from all aspects for their applications and requirements.
TopIntroduction
Corners in digital images give important clues for shape representation and analysis. Since dominant information regarding shape is usually available at the corners, they provide important features for object recognition, shape representation and image interpretation. Corners are the robust features in the sense that they provide important information regarding objects under translation, rotation and scale change. If the corner points are identified properly, a shape can be represented in an efficient and compact way with sufficient accuracy in many shape analysis problem.
Corner points represent important features of an object that may be useful at subsequent level of computer vision. Guru at el (Guru at el., 2004) says that information about a shape is concentrated at the corners and corners practically prove to be descriptive primitives in shape representation and image interpretation. Asada and Brady (Asada and Brady, 1986) insist that these points play dominant role in shape perception by humans. Attneave (Attneave, 1954) proposed that information along a visual contour is concentrated in the regions of high magnitude of curvature. Corner points are used in various computer vision, computer graphics, and pattern recognition applications. It can be used as a step in document image analysis, such as chart and diagram processing (Kasturi et el., 1990) and is also important from the view point of understanding human perception of objects (Attneave, 1954). It plays crucial role in decomposing or describing the curve (Abe et el., 19993). It is also used in scale space theory (Deriche & Giraudon, 1990; Mokhtarian & Mackworth, 1992), image representation (Cabrelli & Molter, 1990), stereo vision (Deriche & Faugeras, 1990, Vincent & Laganire, 2001), motion tracking (Dreschler & Nagel, 1982; Wang & Brady, 1995), image matching (Smith et el., 1998; Vincent & Laganiere, 2005), building 2D mosaics (Zoghlami, et el. 1997) and preprocessing phase of outline capturing systems (Sarfraz et el., 2004a; Sarfraz et el., 2004b).
Corner detection schemes can be broadly divided into two categories based on their applications:
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.
Many corner detection algorithms have been proposed which can be broadly divided into two parts. One is to detect corner points from grayscale images (Harris & Stephens 1988; Kitchen & Rosenfeld, 1982; Noble, 1988; Smith & Brady, 1995) and other relates to boundary based corner detection (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). This chapter mainly deals with techniques adopted for later approach.