China
Africa
Explore chapters and articles related to this topic
Digital Image Fundamentals
Published in Sheila Anand, L. Priya, A Guide for Machine Vision in Quality Control, 2019
Bicubic interpolation is an extension of cubic interpolation for interpolating data points on a two-dimensional regular grid. The interpolated surface is smoother than corresponding surfaces obtained by bilinear interpolation or nearest-neighbor interpolation. Bicubic interpolation uses 16 nearest neighbors instead of 8 nearest neighbors of a point. The intensity value V(x, y) of the unknown point from the known (x, y) is obtained using the following equation: V(x,y)=∑I=03∑i=03aijxiyj
Image-Processing Algorithms
Published in Junichi Nakamura, Image Sensors and Signal Processing for Digital Still Cameras, 2017
When used for two-dimensional interpolation, this is called bicubic interpolation. Figure 8.22 shows a bicubic interpolation algorithm that uses 16 surrounding points to estimate the value of pixel Q. Equation 8.14 shows one of the cubic interpolation functions called a cubic spline. The bicubic interpolation function gives better image quality than the two methods described earlier, but at the cost of much higher computational resource requirements than nearest and linear interpolations.
Basic Concepts of Laser Imaging
Published in Helmut H. Telle, Ángel González Ureña, Laser Spectroscopy and Laser Imaging, 2018
Helmut H. Telle, Ángel González Ureña
Bicubic algorithms: Bicubic interpolation is an extension of one-dimensional cubic interpolation, for interpolating data values from a two-dimensional regular pixel grid. The resulting, interpolated surface is much smoother than corresponding surfaces obtained by bilinear interpolation or nearest-neighbor interpolation (both may yield “ragged” edges). Bicubic interpolation is based on algorithms incorporating Lagrange polynomials, cubic splines, or cubic convolution.
Development of weighted ensemble transfer learning for tomato leaf disease classification solving low resolution problems
Published in The Imaging Science Journal, 2023
Alampally Sreedevi, Chiranjeevi Manike
Bicubic Interpolation method: Bicubic interpolation is a two-dimensional system that helps to enlarge or sharpen the edges in digital images. Here, the low-resolution images are generated with the help of the downsampling of high-resolution images using bicubic interpolation. Moreover, the image pixel gets distorted from one grid to another in interpolation. Consequently, it takes a huge amount of time to process during the resampling of the image. The input image to be applied to the bicubic interpolation method is . It is the modified version of the bilinear interpolation method. Here, a totally of sixteen nearest neighbor values are considered for selecting the pixel coefficients. The intensity value for the point is calculated using the below Equation (5). In addition, the bicubic interpolation method effectively preserves the fine details of a natural image with high contrast. The resultant image obtained from the bicubic interpolation method is .
Uncertainty quantification in digital image correlation for experimental evaluation of deep learning based damage diagnostic
Published in Structure and Infrastructure Engineering, 2021
Nur Sila Gulgec, Martin Takáč, Shamim N. Pakzad
Digital image correlation techniques aim to measure the deformations of the specimen from the images taken by digital cameras. Cameras first capture a reference image in an original unloaded state. They continue taking more pictures as the specimen deforms (Desai, 2016). The reference area in the image, which is called the region of interest (ROI), is divided into square image fragments (i.e. subsets or facets) which have a unique grey-scale pattern of pixels (Figure 2). Each subset needs to be distinct enough to facilitate matching; therefore, the test specimen surface needs to have random gray intensities and deform with the specimen with the loading. The subsets in the reference and other images are matched with the facet matching process (Pan, Xie, & Wang, 2010). Such procedure finds the maximum similarity between the deformed subset centred at point and reference subset centred at P(x, y). Similarity can be found by using different subpixel interpolations, e.g. bilinear interpolation, bicubic interpolation and spline interpolation (GOM, 2013). After the matching process of all subsets in the images, the displacement field computed based on the centre points of the subsets. Strain values are determined by calculating the gradients of the displacement.
Saliency-based segmentation of dermoscopic images using colour information
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2022
Step A.1 – Size/Colour Reduction. The image size is reduced so that the maximum between the number of rows and the number of columns is equal to a value fixed by the user, say Maxdim. Downsampling is done using bicubic interpolation. The reduction of the number of colours, namely colnum, specified by the user, is obtained by a Colour Quantisation (CQ) method. Four different CQ methods are available for this purpose: (Dekker 1994; Bruni et al. 2015, 2017; Ramella and Sanniti Di Baja Ramella and Sanniti Di Baja, 2016a).