CVGIP: Image Understanding最新文献

This paper presents a bibliography of nearly 1300 references related to computer vision and image analysis, arranged by subject matter. The topics covered include computational techniques; feature detection and segmentation; image analysis; two-dimensional shape; pattern; color and texture; matching and stereo; three-dimensional recovery and analysis; three-dimensional shape; and motion. A few references are also given on related topics, such as geometry, graphics, coding and processing, sensors and optical processing, visual perception, neural nets, pattern recognition, and artificial intelligence, as well as on applications.

In this paper we describe optimal processor-time parallel algorithms for set operations such as union, intersection, comparison on quadtrees. The algorithms presented in this paper run in O(log N) time using N/log N processors on a shared memory model of computation that allows concurrent reads or writes. Consequently they allow us to achieve optimal speedups using any number of processors up to N/log N. The approach can also be used to derive optimal processor-time parallel algorithms for weaker models of parallel computation.

How can one identify specularities when an object is illuminated by light that varies in spectrum with direction from the surface? A linear model of color shading can answer this question and also recover surface orientation in non-specular regions using only a single color image of the surface taken under a set of illuminants whose positions, strengths, and spectral content need not be known a priori. The shape-from-color method is based on a Lambertian model. For such a reflectance model the surface normal is related in a linear way to the measured RGB color. Linearity means that the Gaussian sphere is transformed into an ellipsoid in color space, and one can solve for the ellipsoid using least squares; surface normals are recovered only up to an overall orthogonal transformation unless additional constraints are employed. When specularities are present, the least-squares method no longer works. If, however, one views specularities as outliers to the underlying color ellipsoid, then a robust method can still find that surface in RGB space. Here a least-median-of-squares method is used to recover shape and detect specularities at the same time.

In this paper, we analytically characterize the domain of admissible camera locations, orientations, and optical settings for which features of interest in a scene are in focus, inside the field-of-view, and magnified to a certain specification. A general 3D viewing geometry is considered and the camera lens is modeled by a general thick lens model. The analytical relationships describing the complete viewpoint loci that satisfy the above optical feature detectability constraints of resolution, focus, and field-of-view are obtained. These analytical relationships are used as sensor placement constraints in the MVP model-based vision sensor planning system that we have developed. MVP automatically determines viewpoints that satisfy viewing constraints such as the feature detectability constraints discussed in this paper.

A multiresolution approach to curve extraction in images is described. Based on a piecewise linear representation of curves, the scheme combines an efficient method of extracting line segments with a grouping process to identify curve traces. The line segments correspond to linear features defined at appropriate spatial resolutions within a quadtree structure and are extracted using a hierarchical decision process based on frequency domain properties. Implementation is achieved through the use of the multiresolution Fourier transform, a linear transform providing spatially localized estimates of the frequency spectrum over multiple scales. The scheme is simple to implement and computationally inexpensive, and results of experiments performed on natural images demonstrate that its performance compares favorably with that of existing methods.

This paper studies the statistical behavior of errors involved in fundamental geometric computations. We first present a statistical model of noise in terms of the covariance matrix of the N-vector. Using this model, we compute the covariance matrices of N-vectors of lines and their intersections. Then, we determine the optimal weights for the least-squares optimization and compute the covariance matrix of the resulting optimal estimate. The result is then applied to line fitting to edges and computation of vanishing points and focuses of expansion. We also point out that statistical biases exist in such computations and present a scheme called renormalization, which iteratively removes the bias by automatically adjusting to noise without knowing noise characteristics. Random number simulations are conducted to confirm our analysis.