China
Africa
Explore chapters and articles related to this topic
Glossary of Computer Vision Terms
Published in Edward R. Dougherty, Digital Image Processing Methods, 2020
Robert M. Haralick, Linda G. Shapiro
Raster scan order refers to the sequence of pixel locations obtained by scanning the spatial domain of an image in a left to right scan of each image row with the rows taken in a top to bottom ordering. Frame format video images are images scanned in raster scan order.
A novel resolution independent gradient edge predictor for lossless compression of medical image sequences
Published in International Journal of Computers and Applications, 2021
Urvashi Sharma, Meenakshi Sood, Emjee Puthooran
Volumetric medical datasets containing large numbers of image frames are processed by the proposed RIGED predictor on a frame by frame basis. Causal template contains neighboring pixels of an image to be predicted. Pixel prediction is done in raster scan order after extraction of neighboring pixels. RIGED predicts the current pixel at an optimal threshold value for 8- and 16-bit depth images. Residual image is obtained after RIGED prediction by subtracting predicted image from the original image taken from volumetric dataset. Obtained residual has lower entropy value that is further entropy encoded. Entire process of the proposed algorithm is repeated for each frame resulting in efficient prediction measured in terms of entropy. An exhaustive empirical work has been done on large datasets of different resolutions and of different modalities containing varying number of image frames. The novelty of this research paper is to come up with an optimal threshold value for prediction of 8- and 16-bit images and to get the residual image with the lowest entropy. The proposed RIGED is resolution and modality independent.
Recursive Information Hiding Scheme Through LSB, PVD Shift, and MPE
Published in IETE Technical Review, 2018
Mehdi Hussain, Ainuddin Wahid Abdul Wahab, Noman Javed, Ki-Hyun Jung
Hong et al. [8] introduced the prediction error-based reversible data embedding method. In this work, the base pixel () is predicted based on neighboring pixels in a raster scan order. In Figure 2, is predicted by considering previous, , and pixels. The prediction value is calculated by Equation (4), where and represent the maximum and minimum pixel values of and , respectively.
A Monitoring and Diagnostic Approach for Stochastic Textured Surfaces
Published in Technometrics, 2018
Suppose an image is comprised of M pixels, and let Yj = [yj, 1, yj, 2, …, yj, M ]T (j = 1, 2,…, N) denote the set of ordered pixels for the jth image in a sample of N images. We use the subscript j later for indexing images; however, we will often omit it for simplicity, unless necessary. Suppose the elements of Y are ordered in a row raster scan pixel sequence of left-to-right, moving from the top row to the bottom row of the image, as illustrated in Figure 2. Let f(Y) denote the joint distribution of Y, which theoretically provides the most complete characterization of the statistical behavior of the stochastic textured surface. However, it is clearly infeasible to estimate such high-dimensional nonparametric distributions directly. In light of this, consider the factorization: where Y(i) = {yk: k = 1,…, i−1}. The notation is illustrated in Figure 2.