A coarse-to-fine quasi-physical optimization method for solving the circle packing problem with equilibrium constraints

This paper addresses an important extension of the circle packing problem (CPP), the circle packing problem with equilibrium constraints (CPPEC). It considers the dense packing of n circular disks in a large circular container at the same time satisfying the equilibrium constraints. Under the industrial background of the layout design on satellite modules, this NP-hard global optimization problem is important in both theory and practice. We introduce two new quasi-physical models for solving CPPEC in this paper. One is to mimic the elastic movement driven by repelling forces from extruded disks, the other is to simulate a whole translation movement of the disks driven by a pulling force from an imaginative elastic rope connecting the centroid of the disks and the center of the container. Then, inspired by the coarse-to-fine control strategy in the manufacture industry, we propose a coarse-to-fine quasi-physical (CFQP) optimization method that adopts the two quasi-physical models for the quasi-physical descent procedure and combines a basin hopping with tabu method for the search procedure. In this way, not only could CFQP take into account the diversity of the search space to facilitate the global search, but it also does fine search to find the corresponding local minimum in a promising local area. Experiments were on two sets of 11 representative test instances. Computational results showed that CFQP achieved new and better results on four instances, at the same time it matched the current best records on the other six (accurate to 0.0001). Moreover, CFQP resulted in smaller equilibrium deviations than that of others published in the literature. In addition, we generated 34 new CPPEC instances basing on the CPP benchmarks, and provided computational results on the two sets of 34 new CPPEC instances, and the container radii obtained are close to the published results on CPP.     

A novel two-phase Hilbert-scan-based search algorithm for block motion estimation using CTF data structure

Motion estimation is one of the kernel issues in MPEG series. In this correspondence, a novel two-phase Hilbert-scan-based search algorithm for block motion estimation is presented. First in the intra-phase, a segmentation of the Hilbert curve is applied to the current block, then a novel coarse-to-fine data structure is developed to eliminate the impossible reference blocks in the search window of the reference frame. In the inter-phase, a new prediction scheme for estimating the initial motion vector of the current block is presented. Experimental results reveal that when compared to the GAPD algorithm, our proposed algorithm has better execution time and estimation accuracy performance. Under the same estimation accuracy, our proposed algorithm has better execution time performance when compared to the FS algorithm. In addition, when comparing with the TSS algorithm, our proposed algorithm has better estimation accuracy performance, but has worse execution time performance.     

动态场景监控系统中人数统计算法的研究

随着视频监控系统智能化程度的提高,监控区域的人数统计也成为一个热点.提出了一种快速有效的人数统计方法.使用背景差分与帧差分相结合的方式提取运动目标,使用两步分割法区分人群与单个行人,使用基于目标区域算法统计人群的数量的.与目前已有的方法相比,该方法在保持较高准确率的前提下,实时性有了很大的提高,并且降低了对硬件的需求.     

Smooth coverage path planning and control of mobile robots based on high-resolution grid map representation

This paper presents a new approach to a time and energy efficient online complete coverage solution for a mobile robot. While most conventional approaches strive to reduce path overlaps, this work focuses on smoothing the coverage path to reduce accelerations and yet to increase the average velocity for faster coverage. The proposed algorithm adopts a high-resolution grid map representation to reduce directional constraints on path generation. Here, the free space is covered by three independent behaviors: spiral path tracking, wall following control, and virtual wall path tracking. Regarding the covered region as a virtual wall, all the three behaviors adopt a common strategy of following the (physical or virtual) wall or obstacle boundaries for close coverage. Wall following is executed by a sensor-based reactive path planning control process, whereas the spiral (filling) path and virtual wall path are first modeled by their relevant parametric curves and then tracked via dynamic feedback linearization. For complete coverage, these independent behaviors are linked through a new path linking strategy, called a coarse-to-fine constrained inverse distance transform (CFCIDT). CFCIDT reduces the computational cost compared to the conventional constrained inverse distance transform (CIDT), which applies a region growing starting from the current robot position to find the nearest unexplored cell as well as the shortest path to it while constraining the search space. As for experimental validation, performance of the proposed algorithm is compared to those of conventional coverage techniques to demonstrate its completeness of coverage, energy and time efficiency, and robustness to the environment shape or the initial robot pose.     

一种立体匹配的新算法

提出了一种基于遗传算法进行立体匹配，并获得高密度视差图的方法。遗传算法比其它优化方法具有更强的搜索能力，因此非常适用于求解全局优化问题。利用这一特点，该文将每一幅视差图看作是一个个体，将视差值编码形成染色体，这样每个个体中都有许多染色体。然后，引入匹配约束条件，形成目标函数，利用遗传算法搜索全局最优解。另外，还采用了由粗到精的层次策略来减少匹配模糊性和搜索时间。最后，分别用合成图像和真实图像对这种方法进行了验证。     

Coarse-to-fine adaptive masks for appearance matching of occluded scenes

In this paper, we discuss an appearance-matching approach to the difficult problem of interpreting color scenes containing occluded objects. We have explored the use of an iterative, coarse-to-fine sum-squared-error method that uses information from hypothesized occlusion events to perform run-time modification of scene-to-template similarity measures. These adjustments are performed by using a binary mask to adaptively exclude regions of the template image from the squared-error computation. At each iteration higher resolution scene data as well as information derived from the occluding interactions between multiple object hypotheses are used to adjust these masks. We present results which demonstrate that such a technique is reasonably robust over a large database of color test scenes containing objects at a variety of scales, and tolerates minor 3D object rotations and global illumination variations. Received: 21 November 1996 / Accepted: 14 October 1997     

Learning Layered Motion Segmentations of Video

We present an unsupervised approach for learning a layered representation of a scene from a video for motion segmentation. Our method is applicable to any video containing piecewise parametric motion. The learnt model is a composition of layers, which consist of one or more segments. The shape of each segment is represented using a binary matte and its appearance is given by the rgb value for each point belonging to the matte. Included in the model are the effects of image projection, lighting, and motion blur. Furthermore, spatial continuity is explicitly modeled resulting in contiguous segments. Unlike previous approaches, our method does not use reference frame(s) for initialization. The two main contributions of our method are: (i) A novel algorithm for obtaining the initial estimate of the model by dividing the scene into rigidly moving components using efficient loopy belief propagation; and (ii) Refining the initial estimate using α β-swap and α-expansion algorithms, which guarantee a strong local minima. Results are presented on several classes of objects with different types of camera motion, e.g. videos of a human walking shot with static or translating cameras. We compare our method with the state of the art and demonstrate significant improvements.     

S&CNet: A lightweight network for fast and accurate depth completion

Dense depth completion is essential for autonomous driving and robotic navigation. Existing methods focused on attaining higher accuracy of the estimated depth, which comes at the price of increasing complexity and cannot be well applied in a real-time system. In this paper, a coarse-to-fine and lightweight network (S&CNet) is proposed for dense depth completion to reduce the computational complexity with negligible sacrifice on accuracy. A dual-stream attention module (S&C enhancer) is proposed according to a new finding of deep neural network-based depth completion, which can capture both the spatial-wise and channel-wise global-range information of extracted features efficiently. Then it is plugged between the encoder and decoder of the coarse estimation network so as to improve the performance. The experiments on KITTI dataset demonstrate that the proposed approach achieves competitive result with respect to state-of-the-art works but via an almost four times faster speed. The S&C enhancer can also be easily plugged into other existing works to boost their performances significantly with negligible additional computations.     

CLA-GAN: A Context and Lightness Aware Generative Adversarial Network for Shadow Removal

In this paper, we propose a novel context and lightness aware Generative Adversarial Network (CLA-GAN) framework for shadow removal, which refines a coarse result to a final shadow removal result in a coarse-to-fine fashion. At the refinement stage, we first obtain a lightness map using an encoder-decoder structure. With the lightness map and the coarse result as the inputs, the following encoder-decoder tries to refine the final result. Specifically, different from current methods restricted pixel-based features from shadow images, we embed a context-aware module into the refinement stage, which exploits patch-based features. The embedded module transfers features from non-shadow regions to shadow regions to ensure the consistency in appearance in the recovered shadow-free images. Since we consider pathces, the module can additionally enhance the spatial association and continuity around neighboring pixels. To make the model pay more attention to shadow regions during training, we use dynamic weights in the loss function. Moreover, we augment the inputs of the discriminator by rotating images in different degrees and use rotation adversarial loss during training, which can make the discriminator more stable and robust. Extensive experiments demonstrate the validity of the components in our CLA-GAN framework. Quantitative evaluation on different shadow datasets clearly shows the advantages of our CLA-GAN over the state-of-the-art methods.     

Efficient facial landmark localization using spatial–contextual AdaBoost algorithm

Facial landmark detectors can be categorized into global and local detectors. Global facial landmark detectors rely on global statistical relations between landmarks, but do not sufficiently utilize local appearance information, whereas local detectors mainly focus on local appearance attributes of landmarks. Although the AdaBoost algorithm has been successfully employed in object localization, it cannot take advantage of geometric facial feature distribution very well. We propose an AdaBoost algorithm called SC-AdaBoost, which efficiently combines the global knowledge of landmark distribution, the regional shape model, and the local landmark attributes based on a coarse-to-fine strategy. The global prior distribution of landmarks is estimated using a face image set with landmark annotations. First, the face region is detected as a rectangular bounding box using a Haar-like feature-based boosting method, and the global distribution of landmarks is used to determine the facial component regions. Facial landmark localization is roughly performed by regional shape modeling. Posteriors of individual weak classifiers are determined by Gabor wavelet analysis at landmark candidate positions constrained by the regional shape model. SC-AdaBoost is established by empirical risk minimization, which decides the weights for the weak classifiers, and is used for the precise localization. The strength of the proposed approach is shown by extensive experiments using standard face datasets.