Affiliations: Department of Instrumentation and Control Engineering, Netaji Subhas Institute of Technology, University of Delhi, New Delhi, India
Correspondence:
[*]
Corresponding author. Navdeep, Department of Instrumentation and Control Engineering, Netaji Subhas Institute of Technology, University of Delhi, New Delhi, India. E-mail: [email protected].
Abstract: This paper presents an improved hyper smoothing function based methodology for efficient edge detection. The main aim of this work is to obtain localized edges of noisy and blurred images without duplicate ones and integrating them into meaningful object boundaries. Therefore, logarithmic hyper-smoothing function is introduced in local binary pattern leading to improved hyperfunction based local binary pattern (IHLBP) algorithm. The proposed technique uses an improved counting scheme to correctly evaluate the number of image points having pixel value greater than or equal to the central pixel. The IHLBP algorithm is tested on synthetic images, radiography images, real-life pictures from USC-SIPL and BSDS database. Improved local binary pattern (ILBP), hyper local binary pattern (HLBP), Canny and Sobel methods are also used for comparative analysis. The results reveal that the proposed algorithm performs well on all synthetic and real images in the presence of blur and salt & pepper noise. Thus IHLBP proves to be an effective approach for edge detection in comparison to conventional methods.
Keywords: Edge detection, digital radiography images, real images, noise images
