Improved Moving Least Square-Based Multiple Dimension Decomposition (MDD) Technique for Structural Reliability Analysis

This paper presents the state-of-the-art on different moving least square (MLS)-based dimension decomposition schemes for reliability analysis and demonstrates a modified version for high fidelity applications. The aim is to improve the performance of MLS-based dimension decomposition in terms of accuracy, number of function evaluations and computational time for large-dimensional problems. With this in view, multiple finite difference high dimension model representation (HDMR) scheme is developed. This anchored decomposition is implemented starting from an initial reference point and progressively evolving in successive iterations. Most probable point (MPP) of failure is identified in every iteration and is used as the reference point for the next decomposition until it converges. Hermite polynomials in MLS framework are used between the support points for efficient interpolation. The support points are generated sequentially using multiple sparse grids based on the Clenshaw–Curtis scheme. Once the global response surface is constructed using the support points generated in each iteration, importance sampling is employed for reliability analysis. Six different benchmark problems are solved to show its performance vis-à-vis other methods. Finally, reliability-based design of a composite plate is demonstrated, clearly showing the advantage and superiority of the proposed improvements in MLS-based multiple dimension decomposition (MDD).

DeRPN: Taking a Further Step toward More General Object Detection

Most current detection methods have adopted anchor boxes as regression references. However, the detection performance is sensitive to the setting of the anchor boxes. A proper setting of anchor boxes may vary significantly across different datasets, which severely limits the universality of the detectors. To improve the adaptivity of the detectors, in this paper, we present a novel dimension-decomposition region proposal network (DeRPN) that can perfectly displace the traditional Region Proposal Network (RPN). DeRPN utilizes an anchor string mechanism to independently match object widths and heights, which is conducive to treating variant object shapes. In addition, a novel scale-sensitive loss is designed to address the imbalanced loss computations of different scaled objects, which can avoid the small objects being overwhelmed by larger ones. Comprehensive experiments conducted on both general object detection datasets (Pascal VOC 2007, 2012 and MS COCO) and scene text detection datasets (ICDAR 2013 and COCO-Text) all prove that our DeRPN can significantly outperform RPN. It is worth mentioning that the proposed DeRPN can be employed directly on different models, tasks, and datasets without any modifications of hyperparameters or specialized optimization, which further demonstrates its adaptivity. The code has been released at https://github.com/HCIILAB/DeRPN.