Segmentation of 3D range images using pyramidal data structures

Sabata, Bikash; Arman, Farshid; Aggarwal, J. K.

doi:10.1109/iccv.1990.139614

Cited by 22 publications

(19 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many of these criteria are heuristics [6,8,39,47] and some rely on user-defined thresholds. Others, especially recent ones, are applications of statistical and information-theoretic criteria [7, 9, 17, 28-30, 45, 52, 53].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Model Selection Techniques and Merging Rules for Range Data Segmentation Algorithms

Bubna

Stewart

2000

Computer Vision and Image Understanding

View full text Add to dashboard Cite

mentioning

confidence: 99%

“…1). Merging techniques used in vision are based on empirical heuristics [6,15,39,48] and perform poorly at small discontinuities. Further, in an attempt to avoid the model selection problem, many merging techniques only join fits to the same model [6,15,27,48], potentially limiting the effectiveness of merging.…”

mentioning

confidence: 99%

Model Selection Techniques and Merging Rules for Range Data Segmentation Algorithms

Bubna

Stewart

2000

Computer Vision and Image Understanding

View full text Add to dashboard Cite

“…An extensive research on this topic reveals that there are many recent methodologies addressing this problem [5,6,11,4]. In order to treat human pose in uncontrolled scenarios, an early work used range image for object recognition or modeling [10]. This approach achieved a straighforward solution to the problem of intensity and view changes in RGB images through the representation of 3D structures.…”

Section: Introductionmentioning

confidence: 99%

Graph cuts optimization for multi-limb human segmentation in depth maps

Hernández-Vela

Zlateva

Marinov

et al. 2012

2012 IEEE Conference on Computer Vision and Pattern Recognition

View full text Add to dashboard Cite

show abstract

“…al. [2], when switching from single processor to a SIMD parallel processor. Our exact case analysis will give an accurate estimation of the speedups if the values of the different parameters are known.…”

Section: Discussionmentioning

confidence: 99%

“…al. [2][3] was chosen as the candidate for implementation on a SIMD machine. The first stage of the two stage process involves oversegmenting the image based on zeroth and first, order surface properties.…”

Section: Description Of the Algorithmmentioning

confidence: 99%

Complexity analysis of range image segmentation on MasPar MP-1

Ramesh

Chaudhary²

Proceedings of 36th Midwest Symposium on Circuits and Systems

View full text Add to dashboard Cite

-Many low level vision tasks that are computationally intensive are easily parallelizable. The lack of parallel processing systems, or their prohibitive costs, have prevented the move of vision processing algorithms from single processor systems to multiprocessor systems. With the recent spurt of parallel processing hardware, there is a need to investigate the feasibility of using such machines for some vision algorithms. Speedup is an important factor in determining the feasibility of migration from single processor systems to parallel processors. In this work, we investigate a particular segmentation algorithm and present theoretical speedup results. Our formula can work out numerical speedups by simply plugging in the parameter values. INTRODOCTIONComputer vision tasks require an enormous amount of computation demanding high performance computers for practical, redtime applications. Parallelism appears to be the only economical way to achieve this level of performance. Most of the work in computer vision focuses on images with 2-D data, but pragmatic vision problems require 3-D data which is easily available now.Three dimensional data may be represented by a 3-D matrixof intensity values, f ( i , j, k), where each intensity value represents a property associated with the location ( i , j , k ) . A p rimary .goal (and initial step) of computer vision is to abstract 'Lrelevant'' information from an image. This may involve a process called segmentation that groups a set of homogeneous pixels into regions. Homogeneity can be defined by different criteria depending upon the image modality, Segmentation thus reduces the information content in the image to the most relevant and by defining some features of the segmented regions, computer vision scientists hope to extract just enough information to characterize those regions. Such an abstraction will be helpful in other higher level tasks like object recognition (11 and visualization. Two major approaches to segmentation are the region growing ones and the region splitting ones. In region growing, each pixel is considered in relation to its neighbors and pixels that are "closer" in some distance metric are merged. On the other hand, in region splitting, the whole image is initially considered to be one single region and this region is recursively split into smaller regions. Both these approaches are in general amenable to parallel implementation.In this paper we derive the formulas for the theoretical time complexity and speedups obtainable on the implementation of a segmentation algorithm on thr MasPar SIMD machine. In section we briefly describe the segmentation algorithm. In the following section we discuss the architecture of the SIMD machine, MP-1. In section , we derive the time complexity. section 0.0.2 is the concluding section of the paper. DESCRIPTION OF THE ALGORITHMThe segmentation algorithm of Sabata et. al. [2][3] was chosen as the candidate for implementation on a SIMD machine. The first stage of the two stage process involves oversegmenting t...

show abstract

Segmentation of 3D range images using pyramidal data structures

Cited by 22 publications

References 13 publications

Model Selection Techniques and Merging Rules for Range Data Segmentation Algorithms

Model Selection Techniques and Merging Rules for Range Data Segmentation Algorithms

Graph cuts optimization for multi-limb human segmentation in depth maps

Complexity analysis of range image segmentation on MasPar MP-1

Contact Info

Product

Resources

About