An Advanced Approach to Detect Edges of Digital Images for Image Segmentation

Introduction

The problem of image segmentation is well-known and addressed by many scientists because it plays an important role in the various applications based on the image analysis and the computer vision systems. But, still it is considered being one of the most difficult and challenging tasks in image processing and object recognition, and determines the quality of final results of the image analysis. In the domain of image processing (Acharya & Ray, 2005; Bhabatosh, 2011; Gonzalez & Woods, 2008; Hore et al., 2016; Hore, Chatterjee, Chakraborty, & Shaw, n.d.) and image analysis (Al-amri, Kalyankar, & D., 2010; Huang, Wu, & Fan, 2003; Mansoor et al., 2015; Rampun et al., 2019), the interpretation is frequently required and dependent on the difference between foreground and the background. When the human visual system is mimicked by the computer algorithms, several problems may arise. Segmentation is process that divides an image into its various constituent regions or separately highlight different objects. The number of segments or the depth of the hierarchical division is dependent on the subject. That is, segmentation process must stop when the desired segments are achieved or the constituting of objects are separated (Chakraborty, Roy, & Hore, 2016; Palus, n.d.). The process of segmentation must be carefully performed to avoid over or under segmentation related problems (Barnard, Duygulu, Guru, Gabbur, & Forsyth, 2003; Gao, Mas, Kerle, & Pacheco, 2011).

Edge detection is one of the most primitive tools and frequently used in many of the image processing applications to generate some useful information from the images and often used as the pre-processing step in feature extraction and object recognition (Lyu, Fu, Hu, & Liu, 2019; Madireddy et al., 2019). The edge detection have been used by different real life applications like object recognition (Shin, Goldgof, & Bowyer, 2001), target tracking (Boumediene, Ouamri, & Dahnoun, 2007; Etienne-Cummings, Spiegel, Mueller, & Zhang, 1997; Ould-Dris, Ganoun, & Canals, 2006), segmentation (Kaganami & Beiji, 2009; Xiaohan et al., 1992), structural analysis (Kanitkar, Bharti, & Hivarkar, 2011; Luo, Higgs, & Kowalik, 1996), data compression (Deepu & Lian, 2015; Hsu, 1993), biomedical image analysis (Chakraborty, Mali, Banerjee, et al., 2018; Chakraborty, Mali, Chatterjee, Anand, et al., 2018; S. Chakraborty, Mali, Chatterjee, Banerjee, Roy, et al., 2018; Chakraborty, Mali, Chatterjee, Banerjee, Sah, et al., 2018; Chakraborty, Chatterjee, Ashour, Mali, & Dey, 2017; Chakraborty et al., 2018; Chakraborty, Chatterjee, Das, & Mali, 2020; Shouvik Chakraborty & Mali, 2018; Shouvik Chakraborty, Mali, Chatterjee, Banerjee, Mazumdar, Debnath, et al., 2017; Chakraborty, Mali, Chatterjee, Banerjee, Roy, Dutta, et al., 2017; Liu et al., 2014; Pacelli, Loriga, Taccini, & Paradiso, 2006; Roy et al., 2018), cryptographic analysis (Chakraborty, Seal, Roy, & Mali, 2016; Mali, Chakraborty, & Roy, 2015; Mali, Chakraborty, Seal, & Roy, 2015; Roy, Chakraborty, et al., 2019; Roy, Mali, et al., 2019; Seal, Chakraborty, & Mali, 2017) and also help for pattern matching and reconstruction, such as image reconstruction and so on. Edge detection is an active area of research as it facilitates higher level image analysis. In image processing and computer vision, the edge detection methods are focused to find the different sharp changes in the intensity in the image so that some essence of the physical and geometrical features of objects can be extracted. In general, three broad steps associated with the Edge detection process and the steps are as follows: