This paper describes an edge combination algorithm. Our approach is based on the study of four step edge models (ideal, blurred, pulse and staircase) in scale space. Under these conditions, we show that the use of two...
详细信息
ISBN:
(纸本)0818638826
This paper describes an edge combination algorithm. Our approach is based on the study of four step edge models (ideal, blurred, pulse and staircase) in scale space. Under these conditions, we show that the use of two scales (high and low) are sufficient for good edge detection. We derive a set of rules to combine edge information and give an appropriate algorithm taking into account the origin of false edges and their behavior in scale space.
Geometric invariants are shape descriptors that remain unchanged under geometric transformations such as projection, or change of the viewpoint. Earlier we have developed a new method of obtaining local projective and...
详细信息
ISBN:
(纸本)081863880X
Geometric invariants are shape descriptors that remain unchanged under geometric transformations such as projection, or change of the viewpoint. Earlier we have developed a new method of obtaining local projective and affine invariants for a general curve without any correspondences. Being local, the invariants are much less sensitive to occlusion than global invariants. The invariants computation is based on a canonical method. This consists of defining a canonical coordinate system using intrinsic properties of the shape, independently of the given coordinate system. Since this canonical system is independent of the original one, it is invariant and all quantities defined in it are invariant. Here we present a further development of the method to obtain local semi-invariants, that is local invariants for curves with known correspondences. Several configurations are (reeled) curves with known correspondence of one or two feature points or lines.
In this paper, we introduce a new shape modeling scheme that can transform smoothly from local (distributed parameter) to global (lumped parameter) models or vice-versa. The modeling scheme makes use of a hybrid primi...
详细信息
ISBN:
(纸本)081863880X
In this paper, we introduce a new shape modeling scheme that can transform smoothly from local (distributed parameter) to global (lumped parameter) models or vice-versa. The modeling scheme makes use of a hybrid primitive called the deformable superquadric constructed in an orthonormal wavelet bases. These multiresolution bases provide the power to continuously transform from local to global shape deformations and thereby allow for a continuum of shape models - from those with local to those with global shape descriptive power - to be created. The characteristic of continuously transforming from local to global shape deformations allows us to generate fractal surfaces of arbitrary degree that can be useful in describing natural detail. We embed these multiresolution shape models in a probabilistic framework and use them for segmenting anatomical structures in the human brain from MRI data.
This paper proposes a method which produces a 3D structural description from contour data or sparse range data by using superquadrics. The proposed method consists of primal segmentation using super-quadrics and conve...
详细信息
ISBN:
(纸本)081863880X
This paper proposes a method which produces a 3D structural description from contour data or sparse range data by using superquadrics. The proposed method consists of primal segmentation using super-quadrics and convex parts merger using Akaike's information criterion (called AIC) as the criterion. The primal segmentation operation produces many convex parts by expanding superquadrics within the object. However, the AIC criterion enables us to keep the number of parts reasonable because it determines how many parts form the object. An AIC for merging superquadrics is derived in this paper. The proposed method is tested successfully by transforming contour data and 3D sparse range data of human beings into parts descriptions. The tests indicate that the proposed method can be used for 3D object recognition as well as data capture for computer graphics applications.
This paper is concerned with the problem of tracking a moving object having markings on its surface in front of a static camera. In this case the motion analysis reduces to the problem of tracking these markings from ...
详细信息
ISBN:
(纸本)0818638826
This paper is concerned with the problem of tracking a moving object having markings on its surface in front of a static camera. In this case the motion analysis reduces to the problem of tracking these markings from frame to frame. We model the markings by B-splines and by-pass the problem of establishing the correspondence between the image markings points in the various frames by relating the B-spline parameters of the same marking viewed at different frames. Implicit in this relationship is the motion parameters. The complexity of the problem is drastically reduced by first acquiring the 3-D object curve structure before the usual motion estimation.
This paper presents a shape and motion estimation algorithm based on non-linear least squares applied to the tracks of features through time. While our approach requires iteration, it quickly converges to the desired ...
详细信息
ISBN:
(纸本)0818638826
This paper presents a shape and motion estimation algorithm based on non-linear least squares applied to the tracks of features through time. While our approach requires iteration, it quickly converges to the desired solution, even in the absence of a priori knowledge about the shape or motion. Important features of the algorithm include its ability to handle partial point tracks and true perspective, to use line segment matches and point matches simultaneously, and its use of an object-centered representation for faster and more accurate structure and motion recovery.
A system capable of labeling the connected components in multi-valued images at video rates is described. The technique uses a simple linear systolic array to modify the labels of recently visited pixels in a raster s...
详细信息
ISBN:
(纸本)0818638826
A system capable of labeling the connected components in multi-valued images at video rates is described. The technique uses a simple linear systolic array to modify the labels of recently visited pixels in a raster scan. The array removes the need for the complex processing.of label equivalence tables used in previous algorithms. A two-pass system has been designed using the same custom VLSI chip for both passes with minimal support circuitry. This has an advantage over previous hardware implementations where more complex algorithms require separate hardware modules for each pass.
We present a class of region growing algorithms based on an analogy to a mass suspended in a field of forces. The mass represents the growing region, and the field represents the degree of concordance between local pi...
详细信息
ISBN:
(纸本)0818638826
We present a class of region growing algorithms based on an analogy to a mass suspended in a field of forces. The mass represents the growing region, and the field represents the degree of concordance between local pixel values and the color characteristics of the region. The algorithms are particularly well suited to systems that are looking for specific simply-connected shapes in local areas of interest, for example, a rectangle of unknown size, elongation, or orientation. Comparing pseudophysical region growing with Hough transforms, which are also well suited to finding shapes, we show that region growing has competitive space and time complexity, and does not require image preprocessing.for boundaries.
We present a method to correct an approximate polygonal sketch of an image object boundary by adjusting the location of each corner point in the sketch. The method is suitable for the segmentation of unknown images us...
详细信息
ISBN:
(纸本)0818638826
We present a method to correct an approximate polygonal sketch of an image object boundary by adjusting the location of each corner point in the sketch. The method is suitable for the segmentation of unknown images using interactive techniques, and for locating object boundaries in model based segmentation. For each point p at which two lines meet in an open or closed polygonal sketch, a corner segmentation model is derived based on the angle, orientation and scale of the corner defined, and the image function f(x,y) in the immediate region about p. A corner template is then constructed and matched in a small neighborhood about p thereby providing a corrected polygonal sketch. The segmentation model is correct in 95% of the cases and the corner accurately located.
This paper describes the design and implementation of a Single Instruction Multiple Data (SIMD) depth-from-motion algorithm on the image Understanding Architecture Simulator. Correspondences are established in paralle...
详细信息
ISBN:
(纸本)081863880X
This paper describes the design and implementation of a Single Instruction Multiple Data (SIMD) depth-from-motion algorithm on the image Understanding Architecture Simulator. Correspondences are established in parallel for two temporally separated images through correlation The correspondences are used to determine the translational and rotational motion parameters of the camera through a parallel motion algorithm. This is done by first determining the approximate translational parameters and then constraining the search for the exact translational and rotational parameters. Finally the dense depth map is computed from the image correspondences and the computed motion parameters. Results are analyzed for three image sequences acquired from mobile vehicles (the Autonomous Land Vehicle, the Carnegie-Mellon NAVLAB, and the UMass Denning Robot). Depths are obtained at an average accuracy of about 8% in outdoor image sequences. The depth maps are processed to locate relatively small obstacles like cans and cones to a distance of about 60 feet. Larger obstacles like hills are located even when they are much farther away. Finally, excellent results of 3-D surface reconstruction at considerable distances (60-feet) for small objects are reported. Issues related to the speedup and accuracy of the computationally intensive problem of motion analysis are explored in the context of the algorithm.
暂无评论