Gray-level corner detection

Besides reducing thick responses to thin, the application of non-maximum suppression to digital gradient magnitudes also improves the form of the edge response histogram, making the choice of thresholds easier.     

Personal computer corner

This column will be devoted to reporting significant developments in personal computer hardware and software, and covering particularly innovative or effective uses of personal computers in the environmental field. Input and feedback from the readers is essential to achieving these goals.     

Early jump-out corner detectors

The authors present two new corner detectors. One works by using dissimilarity, along the contour direction to detect curves in the image contour, and the other estimates image curvature along the contour direction. These operators are fast, robust to noise, and require no subjective thresholding     

Parallel algorithms for corner stitching

Corner stitching is the underlying data structure that is used to represent rectangular objects in interactive VLSI layout editing systems such as Magic and Tailor. In this paper we develop efficient algorithms for basic corner stitching operations under the message-passing paradigm. These algorithms were implemented using C and PVM on a distributed network composed of SUN workstations. Experimental results show that significant speed-ups were obtained. © 1998 John Wiley & Sons, Ltd.     

一种改进的Harris角点检测算法

图像拼接中,提高角点检测的精确度可以提高配准的精度。在Harris算法的基础上,提出了一种改进的角点检测算法,算法首先分别对图像中每行、每列上所有像素点的Rharris进行X、Y方向的曲线拟合,然后对拟合后的曲线进行分析,若某个像素点的Rharris值在两个方向上都处于“波峰”位置,则将该像素点检测为角点。实验结果表明该算法可以避免阈值的选择,有效地克服了阈值选择不当造成的角点冗余或丢失,提高了角点检测的精度。     

Gradual-SURF算子的角点提取

SURF算子为了改善SIFT的计算复杂度高的问题,简化和近似了DoH（Determinant of Hessian）,这样不仅保证了算法结果的稳定性,也提高了计算效率。但是SURF这样的近似简化过程,损失了图像中的一些渐变信息。对SURF算子进行了改进,在其处理过程中加入了渐变的信息。实验结果表明,提出的G-SURF（Gradual-SURF）算子可以获得更稳定的效果,并且同时计算复杂度也有所改善。     

Algorithms for corner stitched data-structures

Corner Stitching was first presented by Ousterhout as a data-structure for VLSI CAD. This paper describes the data-structure in detail. It presents, in greater depth than previously, algorithms for the basic operations described in Ousterhout's original paper. The algorithms for enumerating and updating arbitrary rectangular areas are new. Their constant space complexity bounds are an improvement over previous algorithms for those operations that were recursive. From a practical standpoint, the elimination of the recursion has also made them much faster.This research was supported by the Australian Government Postgraduate Research Awards Scheme and Xerox Corporation.     

Algorithms for corner stitched data-structures

Corner Stitching was first presented by Ousterhout as a data-structure for VLSI CAD. This paper describes the data-structure in detail. It presents, in greater depth than previously, algorithms for the basic operations described in Ousterhout's original paper. The algorithms for enumerating and updating arbitrary rectangular areas are new. Their constant space complexity bounds are an improvement over previous algorithms for those operations that were recursive. From a practical standpoint, the elimination of the recursion has also made them much faster.     

改进的曲率尺度空间角点检测

在曲率尺度空间（CSS）角点检测中,通常采用高斯滤波对图像进行去噪处理。但该方法很大程度上降低了角点检测的准确性和精确性,因此,提出了利用双边滤波来改进的CSS角点检测算法,可以在很好去噪的同时又保证角点检测的高精度。实验结果也充分验证了该算法较其他算法具有的高效性和准确性。     

多尺度模糊加权角点检测新算法研究

为了提高图像角点检测的准确度和降低噪声对检测效果的影响,将多尺度思想和模糊理论引入到角点检测过程中,在建立了像素点属于角点的隶属度函数的基础上提出一种多尺度模糊加权角点检测新算法.首先将原始图像使用高斯核函数进行变换生成一组响应图像,并将其进行加权叠加得到原始图像的平均角点响应值;再选取合适的阚值进行相关处理得到最终的角点.实验结果表明,该算法不但抗噪性能较好,而且提取出来的角点也较准确.     

A geometric look at corner cutting

In this paper we consider conrner cutting and convexity preserving interpolatory refinement schemes in the plane and on the sphere. Using well-known facts from projective geometry we present a unified approach to such schemes and geometric derivations of simple conditions which guarantee that a scheme generates C¹-curves. Our results generalize all of the results known so far and provide the ground for a new convexity preserving C¹-interpolation scheme with a simple direct shape handle.     

Localization properties of direct corner detectors

In the past, several approaches for directly determining corners in gray-value images have been introduced. The accuracy of an approach has usually been demonstrated experimentally by comparing its results with those obtained by previous schemes. In this contribution we analyze localization properties of existing direct corner detectors by using an analytical model of gray-value corners. For the different approaches we derive implicit equations constraining the corner points and numerically evaluate their locations. Since a gray-value corner is generally defined as the curvature extremum along the edge line, we also compute this position and take it as the reference location for a comparison of the investigated approaches.This work was supported by the Deutsche Forschungsgemeinschaft (DFG) and by the European Union (EU), ESPRIT-Project VIVA (Viewpoint Invariant Visual Acquisition).     

Enhanced low-complexity pruning for corner detection

Low-complexity corner detection is essential for many real-time computer vision applications that need to be executed on low-cost/low-power embedded platforms such as robots. The widely used Shi–Tomasi and Harris corner detectors become prohibitive in such platforms due to their high computational complexity, which is attributed to the need to apply a complex corner measure on the entire image. In this paper, we introduce a novel and computationally efficient technique to accelerate the Shi–Tomasi and Harris corner detectors. The proposed technique consists of two steps. In the first step, the complex corner measure is replaced with simple approximations to quickly prune away non-corners. In the second step, the complex corner measure is applied to a small corner candidate set obtained after pruning. Evaluations using standard image benchmarks show that the proposed pruning technique achieves up to 75 % speedup on the Nios-II platform, while yielding corners with comparable or better accuracy than the conventional Shi–Tomasi and Harris detectors.