China
Africa
Explore chapters and articles related to this topic
Further Discussion and Summary on 2-D Motion Estimation
Published in Yun Q. Shi, Huifang Sun, for Multimedia Engineering, 2017
Block matching certainly belongs to the region-based approach (here region means a rectangular block). For an original block in a (current) frame, block matching searches for its best match in another (previous) frame among candidates. Several dissimilarity measures are utilized, among which the mean absolute difference (MAD) is used most often.
Further Discussion and Summary on 2-D Motion Estimation
Published in Yun-Qing Shi, Huifang Sun, Image and Video Compression for Multimedia Engineering, 2019
Block matching certainly belongs to the region-based approach. By region, we mean a rectangle block. For an original block in a (current) frame, block matching searches for its best match in another (previous) frame among candidates. Several dissimilarity measures are utilized, among which the mean absolute difference (MAD) is used most often.
Quantifying the relationships between network distance and straight-line distance: applications in spatial bias correction
Published in Annals of GIS, 2021
We then evaluate the accuracy of DI prediction based on the k-nearest neighbour (kNN) regression with a single variable of mean Euclidean distance of urban patches. kNN is an intuitive and efficient classification method in pattern recognition. An object is classified by a voting of its neighbours in the feature space. The object is assigned the most common class among its k nearest neighbours. kNN regression is used for estimating continuous variables according to similar rules. In kNN regression, the voting result is the average (or weighted average) of the property values of its k nearest neighbours (Altman 1992). Leave-one-out cross-validation was used to evaluate the performance of the kNN regression. After multiple experiments, k = 10 yielded the most accurate result (see Table A2 for full results). The Mean Absolute Difference is 0.042 and the Mean Absolute Percent Error is 3.123%. To evaluate the impacts of the DI prediction errors, we selected the city with a large error (i.e. Changsha), and compared its K-function outcomes using both the actual and predicted DI. As shown in Figure 10(b), K-function using the predicted DI also reduces most of the bias (overestimation of the clustering patterns) caused by the planar K-function (ED-based) just as the network K-function (ND-based) does.
Contributions to humanitarian logistics
Published in IISE Transactions, 2019
Rajan Batta, Simin Huang, Bahar Kara
The first paper entitled “Inequity-averse shelter location for disaster preparedness,” by M. Mostajabdaveh, W. J. Gutjahr, and F. Sibel Salman studies the problem of selecting a set of shelter locations in preparation for natural disasters. Their modeling framework includes both efficiency and inequity. To achieve this, they minimize a linear combination of: (i) the mean distance between opened shelter locations and the locations of the individuals assigned to them; and (ii) Gini’s Mean Absolute Difference of these distances. A stochastic programming model with a set of scenarios that consider uncertain demand and disruptions in the transportation network is developed. A specialized Genetic Algorithm is developed to solve large problem instances. A case study based on Istanbul data is presented to derive insights for decision-makers.
Separation of Machine-Printed and Handwritten Texts in Noisy Documents using Wavelet Transform
Published in IETE Technical Review, 2019
In order to reduce the computations while calculating BVLCC along with DWT and SDDWT, the formula of correlation for each local window is modified and is described by Equation (9). Here, modification is substitution of mean absolute difference of pixels between two blocks in place of local covariance (numerator) and local variance (denominator) is replaced with mean absolute difference of pixels in a block. The modified local correlation coefficient is given as where m denotes sub-bands and V is the level of 2D-DWT and SDDWT. Δ now represents shift in two directions, i.e. lower and left {(0,−k), (−k,0)}. In Equation (9), νl is mean absolute difference of four end pixels in the block and νl+ Δ(k) is also a mean absolute difference of the block shifted by Δ(k). In Equation (10), νl is given as Modified BVLCC is given as where S2 is shift in two directions {(0,−k), (−k,0)}.