Improved Depth Local Binary Pattern for Edge Detection of Depth Image

Depth images generally have low resolution, which leads to poor edge map from conventional edge detection methods. This necessitates the requirement of a consistent edge detection method for depth image applications. Therefore in this paper, improved depth local binary pattern (IDLBP) edge detection method based on local binary pattern is proposed for depth images. Experimentation is performed on synthetic and depth images to verify the validity of suggested technique. It is revealed that IDLBP is a fast, accurate and non-destructive edge detection technique. The depth local binary pattern, improved local binary pattern, hyper smoothing function based local binary pattern, Canny, Sobel and Laplacian of Gaussian edge detection techniques are also realized for comparative analysis. The effectiveness of suggested IDLBP algorithm is evaluated in terms of accuracy, Jaccard similarity index, specificity, sensitivity and dice similarity coefficient.