Academia.eduAcademia.edu

Outline

Title
Abstract
Discussion: The Scale Factor
The Algorithm and Complexity Analysis
Summary and Conclusions
References
All Topics
Engineering
Signal Processing

Fast Computation of the Difference of Low-Pass Transform

Profile image of Richard SternRichard Stern

1984, IEEE Transactions on Pattern Analysis and Machine Intelligence

https://doi.org/10.1109/TPAMI.1984.4767504
visibility

…

description

44 pages

link

1 file

Abstract

This paper defines the Difference of Low-Pass (DOLP) transform and describes a fast algorithm for its computation. The DOLP is a reversible transform which converts an image into a set of band-pass images. A DOLP transform is shown to require O(N2) multiplies and produce O(N Log(N)) samples from an N sample image. When Gaussian low-pass filters are used, the result is a set of images which have been convolved with difference of Gaussian ( DOG) filters from an exponential set of sizes. A fast computation technique based on "resampling" is described and shown to reduce the DOLP transform complexity to O(N Log(N)) multiplies and O(N) storage locations. A second technique, "cascaded convolution with expansion", is then defined and also shown to reduce the computational cost to O(N Log(N)) multiplies. Combining these two techniques yields an algorithm for a DCLP transform that requires O(N) storage cells and requires O(N) multiplies. The DOLP transform provides a basis for a structural description of gray-scale shape. Descriptions of shape in this representation may be matched efficiently to descriptions of shape from other images to determine motion or stereo correspondence. Such descriptions may also be matched independent of their size or image plane orientation.

Related papers

Stereo Correspondence Based on Multiresolution Analysis
Dibyendu Mukherjee

2010

Disparity estimation from stereo imagery has gained substantial interest of research community from its commencement with the recent trend being the use of multiresolution methods. Existing multiresolution based methods generally rely on approximate band based matching neglecting other subbands that carry high-frequency information. Present research is an effort to find a multiresolution based stereo correspondence method that effectively uses the high and low-frequency subbands in multiple resolutions, bridges the gap between feature-based and area-based matching by incorporating the vast feature space of multiresolution and develops an optimum approach between global and local area-based methods in terms of accuracy, implementation and computational complexity. As a response to the lack of exploration of different multiresolution based stereo, a novel comprehensive comparison framework is proposed to evaluate different multiresolution based disparity estimation methods. Extensive ...

View PDFchevron_right
Matching techniques to compute image motion
Andrea Giachetti

Image and Vision Computing, 2000

This paper describes a thorough analysis of the pattern matching techniques used to compute image motion from a sequence of two or more images. Several correlation/distance measures are tested, and problems in displacement estimation are investigated. As a byproduct of this analysis, several novel techniques are presented which improve the accuracy of ow vector estimation and reduce the computational cost by using lters, multi-scale approach and mask sub-sampling. Furthermore new algorithms to get a sub-pixel accuracy of the ow are proposed. A large amount of experimental tests have been performed to compare all the techniques proposed, in order to understand which are the most useful for practical applications, and the results obtained are very accurate, showing that correlation-based ow computation is suitable for practical and real-time applications.

View PDFchevron_right
RASCor: Realtime Associative Stereo Correspondence
yusuf ozturk

2007

Abstract This paper presents a hardware implementation solution to a real time stereo matching problem using system of associative relations (SOAR) computational model. SOAR makes use of pair-wise pixel interactions captured as direction of derivatives to determine the underlying structure of associations within an analysis token. SOAR also offers a similarity measure to assess similarity of two such structures of associations.

View PDFchevron_right
A fast stereo matching method
Changming Sun

Digital Image Computing: Techniques and Applications, 1997

Stereo matching is important in the area of computer vision and photogrammetry. This paper presents a fast stereo matching algorithm which produces a dense disparity map by using a pyramid structure, fast correlation and dynamic programming techniques. Fast correlation is ...

View PDFchevron_right
Fast 2-D Distance Transformations
Stelios Krinidis

IEEE Transactions on Image Processing, 2012

The performance of a number of image processing methods depends on the output quality of a distance transformation (DT) process. Most of the fast DT methodologies are not accurate, whereas other error-free DT algorithms are not very fast. In this paper, a novel, fast, simple, and error-free DT algorithm is presented. By recording the relative-and-coordinates of the examined image pixels, an optimal algorithm can be developed to achieve the DT of an image correctly and efficiently in constant time without any iteration. Furthermore, the proposed method is general since it can be used by any kind of distance function, leading to accurate image DTs.

View PDFchevron_right
PARALLEL IMPLEMENTATION OF IMAGE MATCHING
Sheena Baburaj

Journal of Information and Computational Science, 2019

The concept of image matching is widely used in the areas of security, medical and computer vision which require comparing two images for similarities. However, depending on the size of images, it is highly possible that the application computation cannot be handled in a single process running a sequential algorithm. In order to overcome this limitation, parallel computing is introduced through the Message Passing Interface (MPI) library. For the comparison of two images, both images are first converted into grayscale and then are compared using the Sum of Square Differences (SSD) algorithm. Further, a parallel network of processors was implemented for image matching and to calculate the performance of the SSD algorithm between both images. The performance gain of processors was compared with the performance of a single core processor. Then, using multiple threads the image matching using SSD is implemented. The performance gain of multiple threads was compared with the performance of a single thread. Finally, performance gain between multiple threads and multi-core processors are compared. The comparison results presented a linear relationship between the performance gain and the number of processors used for execution. Hence, it proves that there are significant benefits of parallelism on SSD applications.

View PDFchevron_right
High level synthesis of stereo matching: Productivity, performance, and software constraints
Deming Chen
View PDFchevron_right
Stereo Correspondence Estimation Using Multiwavelets Scale-Space Representation-Based Multiresolution Analysis
Saeid Nahavandi

Cybernetics and Systems, 2008

View PDFchevron_right
Search Space Reduction in the Edge Based Stereo Correspondence
Payman Moallem

2001

Usually, the stereo correspondence for a feature point in the first image is obtained by searching in a predefined region of the second image, based on the epipolar line and the maximum disparity. The reduction of the search space can increase the performance of the matching process, in the co text of the execution time and the accuracy. For the edge-based stereo correspondence, we obtain the noticeable reduction in the search space. Considering the maximum of the disparity gradient in the real scene, we formulated the relation between the maximum search space in the second images with respect to the relative displacement of the continuous edges (as the feature points) in the successive scan lines of the first images. Then we developed some very fast stereo matching algorithms, based on the normalized cross correlation criteria (NCC) for different sizes of the matching block.

View PDFchevron_right
A Taxonomy of Image Matching Techniques for Stereo Vision
Kurt Kubik

Being as a relatively new approach of signalling, moving-block scheme significantly increases line capacity, especially on congested railways. This paper describes a simulation system for multi-train operation under moving-block signalling scheme. The simulator can be used to calculate minimum headways and safety characteristics under pre-set timetables or headways and different geographic an Abstract Stereo vision is a method of depth perception, in which depth information is inferred from two (or more) images of a scene, taken from di erent perspectives. Applications for stereo vision include aerial photogrammetry, autonomous vehicle guidance, robotics, industrial automation and stereomicroscopy.

View PDFchevron_right

References (19)

  1. l91 Abramntic, J. F. and 0. I). Faugcrns. Scqucntial Convolution 'I-cchniqucs for Image Fittcring. IEEE Trans. o)t Acous. Speech and Sipial Processing ASSP-30( 1): 1-10, February, 1982. Aho, Alfred V., John E. Hopcroft. and Jcffcry D. Ullman. Cornpurer Science arid I) formation Processing: The Design and Aiialysis of Cornpurer Algorithms Addison Wcslcy, Reading ,Massachusetts, 1974. Binford, Thomas 0. Survcy of Modcl-hscd Irnagc Analysis Systems. Roborics Research 1( 1): 18-64, Spring, 1982.
  2. Burt, Peter J. Fast, Hierarchical Correlations wirh Gaussian-Like Kern&. Tcchnical Report TR-860, Computer Vision Laboratory, University of Maryland, January, 1980. Burt, Pcter J. Fast Filter Transforms for Image Processing, Cornpuler Graphics arid Image Processing 16:tO-51,198 1.
  3. Campbell, F. W. The Transmission of Spatial Information through the Visual System.
  4. In F. 0. Schrnitt and F. G > Worden (editor), The Neurosciences: Third Srudy Program, . MIT Press, 1974.
  5. Campbell, F. W. and J. G. Robson. Applications of Fourier Analysis to the Visibility of Gratings. Journal of PhysioIogy :551-566.1971.
  6. Crowley, J. L. and A. C. Parker.
  7. The Analysis, Synthesis, and Evaluation of Local Measures for Discrimination and Segmentation of In Conference on Pariern Recognilion and Image Processing, pages 372-378. IEEE Computer Socicty. Textured Regions. June, 1978.
  8. Crowley, J. L. and A. C. Parker. Transfer Function Analysis of Picture Processing Operitors.
  9. In Robert M. Haralick and J. C. Simon (cditor), Issues In DigilaZ Image Processing, chapter 1, pages 3-30. Sijthoff & Noordhoff, 1980.
  10. Crowley, James L. A Represen ralion for Visual Informalion. PhD thcsis, Carnegie-Mcllon Univcrsity, November, 198 1. Crowlcy, 1. L. A Rcprcscntation for Shapc nascd on Pcaks and Ridges in tlic Diffcrcncc of Low-Pass l'ransfom. To be Subinitred to IEEl; Trotis. on P.A.'\[.I. , 1982.
  11. Hall, E. L., Rougc, Lt. J. and Wong, R. Y. Hicrarchical Scarch for lrnagc Matching. In Proc. 1976 IEEE Cor$ oti Decision and Control. pagcs 791-796. IEEE, Dcccmbcr, 1976. Marr, D. and Poggio, T. A Computational Theory of Human Vision. Proc. R. Soc. Lond B , 1979.
  12. Man, D. L., and Hildrcth, E. Theory of Edge Detection. Technical Report A.I. Memo 518, M.I.T., April, 1979. Moravec, H. P. Obstucle Avoidance and Navigation in the Real World by a Seeing Robot Rover. PhD thcsis, Stanford University, September, 1980. Nyquist, H. Certain Factors Affecting Telegraph Speed. Bell Systems Tech Journal 3(2):324-346, April, 1924.
  13. Oppcnheim, A. V. and Schafcr, R. W. Digital Signal Processing. Prentice-Hall inc., Englcwood Cliffs, N. J., 1975.
  14. Paley R.EA.C. and N. Wiener. Fourier Transfonns in ihe Complex Domain American Mathematical Society Coloquium, 19, New York, 1934. Papoulis, A. Systems Sciences: Systems and Transfonns wiih Applications in Optics McGraw-Hill, New York, 1968.
  15. Pratt, William K. Wiley-Interscience: Digital Image Processing John Wiley & Sons, 1978. page 322.
  16. Rosenfcld A. and Vanderbrug, G. i. Coarse-Fine Template Matching. IEEE Trans on Man. Systems and Cybernetics SMC-7(2):104-107, Feb., 1977.
  17. Sachs. M. , J. Nachinias, and 1. G. Robson. Spatial-l-.rcqucncy CIianncls in Human Vision. Jourtinlof the OpticalSociety of ,.imerica61:1176-1186, 1971.
  18. Thomas. J. P. Spacial Rcsoiution and Spatial Interaction.
  19. In E. C. Cartcrcttc and M. P. Friedman (cditor), Handbook of Perceprion. Vol V: Seeing. . Acadcmic Press, New York. 1975.

Related papers

Fast algorithms for low-level vision
Rachid Deriche

IEEE Transactions on Pattern Analysis and Machine Intelligence, 1990

We propose in this paper a recursive filtering structure that drastically reduces the computational effort required for smoothing, performing the first and second directional derivatives or carrying out the Laplacian of an image. These operations are done with a fixed number of multiplications and additions per output point independently of the size of the neighborhood considered. The key to our approach is first the use of an exponentially based filter family and second the use of the recursive filtering. Applications to edge detection problems and multiresolution techniques are considered and a new edge detector allowing the extraction of zero-crossings of an image with only 14 operations per output element at any resolution is proposed. Various experimental results are shown.

View PDFchevron_right
Sequential Local Transform Algorithms for Gray-Level Distance Transforms
Hicham G Elmongui

In this paper, new algorithms are presented for the calculation of gray-level distance transforms, in which the distance values of pixels are proportional to the grayvalue dierences of minimal paths , not the gray v alues themselves, which is the case in [1] and . The presented algorithms are sequential local transform algorithms. The performance of the algorithms are evaluated by testing how quickly they converge to the errorfree distance image. It is shown that the presented algorithms are faster than the previously presented algorithms [5] for gray-level distance transforms. It turns out that the 4-neighbor, 4-raster algorithm is the fastest, converging to a distance image with no erroneous pixels in only two iteration rounds. Furthermore, because of the raster scanning approach they are easily applied other image grids than the rectangular one, and are easy to implement.

View PDFchevron_right
Sequential Local Transform Algorithms for Gray-Level
Hicham Elmongui

In this paper, new algorithms are presented for the calculation of gray-level distance transforms, in which the distance values of pixels are proportional to the grayvalue dierences of minimal paths [5], not the gray v alues themselves, which is the case in [1] and [3]. The presented algorithms are sequential local transform algorithms. The performance of the algorithms are evaluated by testing how quickly they converge to the errorfree distance image. It is shown that the presented algorithms are faster than the previously presented algorithms [5] for gray-level distance transforms. It turns out that the 4-neighbor, 4-raster algorithm is the fastest, converging to a distance image with no erroneous pixels in only two iteration rounds. Furthermore, because of the raster scanning approach they are easily applied other image grids than the rectangular one, and are easy to implement.

View PDFchevron_right
A General Algorithm for Computing Distance Transforms in Linear Time
Jos B. T. M. Roerdink, Wim H Hesselink

2002

A new general algorithm for computing distance transforms of digital images is presented. The algorithm consists of two phases. Both phases consist of two scans, a forward and a backward scan. The first phase scans the image column-wise, while the second phase scans the image row-wise. Since the computation per row (column) is independent of the computation of other rows (columns), the algorithm can be easily parallelized on shared memory computers. The algorithm can be used for the computation of the exact Euclidean, Manhattan (L 1 norm), and chessboard distance (L ∞ norm) transforms.

View PDFchevron_right
Fast Edge-Based Stereo Matching Algorithm through Search Space Reduction
Javad Haddadnia

Ieice Transactions on Information and Systems, 2002

Finding corresponding edges is considered being the most difficult part of edge-based stereo matching algorithms. Usually, correspondence for a feature point in the first image is obtained by searching in a predefined region of the second image, based on epipolar line and maximum disparity. Reduction of search region can increase performances of the matching process, in the context of execution time and accuracy. Traditionally, hierarchical multiresolution techniques, as the fastest methods are used to decrease the search space and therefore increase the processing speed. Considering maximum of directional derivative of disparity in real scenes, we formulated some relations between maximum search space in the second images with respect to relative displacement of connected edges (as the feature points), in successive scan lines of the first images. Then we proposed a new matching strategy to reduce the search space for edge-based stereo matching algorithms. Afterward, we developed some fast stereo matching algorithms based on the proposed matching strategy and the hierarchical multiresolution techniques. The proposed algorithms have two stages: feature extraction and feature matching. We applied these new algorithms on some stereo images and compared their results with those of some hierarchical multiresolution ones. The execution times of our proposed methods are decreased between 30% to 55%, in the feature matching stage. Moreover, the execution time of the overall algorithms (including the feature extraction and the feature matching) is decreased between 15% to 40% in real scenes. Meanwhile in some cases, the accuracy is increased too. Theoretical investigation and experimental results show that our algorithms have a very good performance with real complex scenes, therefore these new algorithms are very suitable for fast edge-based stereo applications in real scenes like robotic applications. key words: fast edge-based stereo, search spaces reduction, directional derivative of disparity, hierarchical multiresolution trical Engineering from Tehran Polytechnic University as the first rank in June 1973, and his M.S. and Ph.D. degrees in Javad Haddadnia received the BS.C degree in Electrical and Electronic Engineering with the first rank in 1993 and MS.c in Electronic Engineering in 1995, respectively, all from Amirkabir University of Technology, Tehran, Iran. He is now a Ph.D. candidate in the Electrical Engineering Department of Amirkabir University of Technology, Tehran, Iran. His research interests include digital image processing, computer vision and face detection and recognition. He has published many papers in the area of image processinhg. He served as a visiting research at the University of Windsor, Canada. He is a member of SPIE and CIPPR.

View PDFchevron_right

Related topics

  • Information Systems
  • Computer Science
  • Algorithms
  • Image Processing
  • Band Pass Filter
  • Electrical and Electronic Engine...
  • Bandpass Filters
  • Difference of Gaussian
  • Fast Algorithm
  • Band pass filters

    • Cited by

    Adaptive scale fixing for multiscale texture segmentation
    Tardi Tjahjadi

    IEEE Transactions on Image Processing, 2000

    This paper addresses two challenging issues in unsupervised multiscale texture segmentation: determining adequate spatial and feature resolutions for different regions of the image, and utilizing information across different scales/resolutions. The center of a homogeneous texture is analyzed using coarse spatial resolution, and its border is detected using fine spatial resolution so as to locate the boundary accurately. The extraction of texture features is achieved via a multiresolution pyramid. The feature values are integrated across scales/resolutions adaptively. The number of textures is determined automatically using the variance ratio criterion. Experimental results on synthetic and real images demonstrate the improvement in performance of the proposed multiscale scheme over single scale approaches.

    View PDFchevron_right
    580 California St., Suite 400
    San Francisco, CA, 94104
    © 2025 Academia. All rights reserved