The problem we are addressing in Alvey Project MMI149 is that of using computer vision to understand the unconstrained 3D world, in which the viewed scenes will in general contain too wide a diversity of objects for topdown recognition techniques to work. For example, we desire to obtain an understanding of natural scenes, containing roads, buildings, trees, bushes, etc., as typified by the two...

The aim of this paper is to show how image points can be extracted accurately. We will restrict our search to speciic points identiied by corners, which are stable given a sequence. Our approach makes use of a model{based corner detector. It matches a part of the image containing a corner against a predeened corner model. Once the tting is accomplished, the position of the corner in the image c...

In this paper, a role of corner detection in outline capturing system is studied/discussed, which is normally used as phase 1 of capturing. Three outline capturing techniques are studied, one without involving any corner detection and two techniques with most suited corner detectors. One common curve approximation method, with cubic Bezier curves, is used in all above techniques in order to stu...

There are many applications such as stereo matching, motion tracking and image registration that require so called ‘corners’ to be detected across image sequences in a reliable manner. The Harris corner detector is widely used for this purpose. However, the response from the Harris operator, and other corner operators, varies considerably with image contrast. This makes the setting of threshold...

A novel distinguished region detector, complementary to existing approaches like Harris-corner detectors, Difference of Gaussian detectors (DoG) or Maximally Stable Extremal Regions (MSER) is proposed. The basic idea is to find distinguished regions by clusters of interest points. In order to determine the number of clusters we use the concept of maximal stableness across scale. Therefore, the ...

In this paper, a novel corner detection method is presented, to extract geometrically important corners for robot applications, such as indoor mobile robot navigation or manipulation. Intensity‐based corner detectors, such as the Harris corner, can detect corners in noisy environments, but have inaccurate corner positions and miss the corners of obtuse angles. Edge‐based corner det...

This paper shows how canonical correlation can be used to learn a detector for corner orientation invariant to corner angle and intensity. Pairs of images with the same corner orientation but different angle and intensity are used as training samples. Three different image representations; intensity values, products between intensity values, and local orientation are examined. The last represen...

We define a universal gradient-based representation for V-, Tand X-type corners. This expression is used to define a corner detector, based on Moments of the Gradient in Scale-space (MoGS). The performance of this operator is compared with that of the Plessey and the Kitchen and Rosenfeld operators. The MoGS operator is shown to have better invariance to scale, to be more reliable and to provid...

This paper presents a multiscale corner detection method based on continuous wavelet transform (CWT) of contour images. The corner points are detected from the local wavelet transform modulus maxima (WTMM) of the contour orientation. To reduce the side effects of the discretization and smoothing that are introduced by the preprocessing steps, we adopt a simple but efficient post processing algo...

In this paper we introduce a fast, robust, accurate eye localization algorithm. Detecting and normalizing human faces from live video streams is the first crucial step in a face verification/recognition system. The accuracy and robustness affect the performance of the following face registration and classification. To localize face regions properly, we detect eye corners by using a corner detec...