Three-Dimensional Surface Reconstruction on the AT&T Pixel Machine

Many low-level image analysis problems which are posed as variational principles can be numerically solved using local and iterative relaxation algorithms. Because of the structure of these algorithms, processing time will decrease nearly linearly with the addition of processing nodes working in parallel on the problem. In this paper, we discuss the implementation of a low-level three-dimensional surface reconstruction algorithm on the AT&T Pixel system. The Pixel machine contains a multiple instruction / multiple datapath (MIMD) computing array which can be used to significantly reduce the computation time of the surface reconstruction algorithms. The computing array has an 8x8 array of processing nodes. Each node has a DSP32 digital signal processor, a high-speed 32-bit programmable device. as its main processor, 36 kilobytes of program memory and .75 megabytes of data memory. The performance of the implementation is evaluated in terms of the absolute processing time.

[1]  P. J. Denning,et al.  Research questions for performances analysis of supercomputers , 1987 .

[2]  A. N. Tikhonov,et al.  Solutions of ill-posed problems , 1977 .

[3]  Edward J. Delp,et al.  Invariant reconstruction of curves and surfaces with discontinuities with applications in computer vision , 1990 .

[4]  E. Delp,et al.  Invariant recovery of curves in m-dimensional space from sparse data , 1990 .

[5]  Eric L. W. Grimson,et al.  From Images to Surfaces: A Computational Study of the Human Early Visual System , 1981 .

[6]  Demetri Terzopoulos,et al.  The Computation of Visible-Surface Representations , 1988, IEEE Trans. Pattern Anal. Mach. Intell..

[7]  Alexander V. Tikhonov,et al.  Ill-Posed Problems in Natural Sciences , 1989 .

[8]  Edward J. Delp,et al.  Invariant reconstruction of visual surfaces , 1989, [1989] Proceedings. Workshop on Interpretation of 3D Scenes.

[9]  Edward J. Delp,et al.  Viewpoint invariant recovery of visual surfaces from sparse data , 1990, [1990] Proceedings Third International Conference on Computer Vision.

[10]  Demetri Terzopoulos,et al.  Multilevel computational processes for visual surface reconstruction , 1983, Comput. Vis. Graph. Image Process..

[11]  Tomaso Poggio,et al.  Computational vision and regularization theory , 1985, Nature.

[12]  J. Meinguet.  Multivariate interpolation at arbitrary points made simple , 1979 .

[13]  Jean Duchon,et al.  Splines minimizing rotation-invariant semi-norms in Sobolev spaces , 1976, Constructive Theory of Functions of Several Variables.