Y. Aloimonos and D. Shulman, Integration of Visual Modules --- An extension of the Marr Paradigm, Academic Press, London, 1989.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....vision systems can function without them. And in fact, research in computational vision has confirmed that combining multiple cues results in a more accurate and less ambiguous result, and a more robust system; typically the limits of only a few of the cues are taxed at any particular moment [4 8]. Such robustness is the only alternative to controlling the visual environment, and is almost certainly required for the development of a wide range of visual applications, including perceptual user interfaces and multi modal interactive systems. These applications also bring real time ....

....infeasible, thus motivating us to address the second problem mentioned above that of cue selection. Some approaches explicitly describe interactions between visual modules by modeling the flow and priority of information between a variety of modules to specify how various cues are combined [4, 17, 18]. Other methods use robust statistics [19, 7] or voting approaches [8] which take advantage of consistency among those cues which have not failed in selecting between cues or alternative interpretations. No aspect of current approaches for cue integration is geared towards timecritical ....

Aloimonos, Y., Shulman, D.: Integration of Visual Modules: An Extension of the Marr Paradigm. Academic Press (1989)

....definition of intrinsic dimension that realizes its triangular structure in section 3.1. A classification of edge ness or corner ness based on a local image patch without taking the context into account always faces the problem of the high degree of ambiguity of visual information (see, e.g. [1]) Taking into account this ambiguity we do not want to come to a final decision about the junction ness of edge ness of an image patch but we want to associate confidences to such classifications. Assigning confidences instead of binary decisions at low level stages of processing has been ....

....of edge ness of an image patch but we want to associate confidences to such classifications. Assigning confidences instead of binary decisions at low level stages of processing has been proven useful since it allows for stabilizing such local classifications according to the context (see, e.g. [1, 21]) By making use of barycentric coordinates (see, e.g. 7] we will utilize the triangular structure of intrinsic dimension to express confidences for the different possible interpretation in section 3.2. This leads to continuous definition of intrinsic dimensionality that covers i0D, i1D and ....

[Article contains additional citation context not shown here]

J. Aloimonos and D. Shulman. Integration of Visual Modules --- An extension of the Marr Paradigm. Academic Press, London, 1989.

....within a process of recurrent prdictions. 1 Introduction Vision, although widely accepted as the most powerful sensorial modality, faces the problem of an extremely high degree of vagueness and uncertainty in its low level processes such as edge detection, optic ow analysis and stereo estimation [1]. However, by integrating information across visual modalities (see, e.g. 9] the human visual systems acquires visual representations which allows for actions with high precision and certainty within the 3D world even under rather uncontrolled conditions. The power of modality fusion arises ....

....of intrinsic relations given by deterministic and statistical regularities across visual modalities. The essential need for fusion of visual modalities, beside their improvement as isolated methods, has also been recognised by the computer vision community during the last 10 years (see, e.g. [1, 3]) Two important regularities in visual data with distinct properties are motion (most importantly rigid body motion, RBM, see, e.g. 5] and statistical interdependencies between features such as collinearity and symmetry (see, e.g. 23] RBM re ects a geometric dependency in the ....

J. Aloimonos and D. Shulman. Integration of Visual Modules | An extension of the Marr Paradigm. Academic Press, London, 1989.

.... e ect of illusionary contours appears after 5 months [7, 38] In the human visual system apart from local orientation other modalities such as color and optic ow are also computed (see, e.g. 20] All these low level processes face the problem of a high degree of vagueness and uncertainty [1]. This arises from a couple of factors. Some of them are associated with image acquisition and interpretation: Owing to noise in the acquisition process along with the limited resolution of cameras, only inaccurate estimates of semantic information (e.g. orientation) are possible. Furthermore, ....

....within and across visual modalities. The essential need for integrating visual information in addition to optimizing single visual modalities to design ecient arti cial visual systems has also been recognized in the vision community after a long period of work on improving single modalities [1]. Gestalt principles are also a ected by multiple visual modalities. For example, gure 1 shows how grouping based on the Gestalt law collinearity can be intensi ed by the di erent modalities contrast transition, optic ow and color. This paper addresses the statistics of natural images in these ....

[Article contains additional citation context not shown here]

J. Aloimonos and D. Shulman. Integration of Visual Modules | An extension of the Marr Paradigm. Academic Press, London, 1989.

....We would like to point out that it is not our aim to derive a perfect stereo system. Stereo is an ambiguous visual modality since the correspondence problem can become extremely awkward in complex scenes and mismatches lead to wrong 3D estimates. Integration of other visual modalities (see, e.g. [1, 18, 3]) and integration of ambiguous information over time (see, e.g. 4, 13, 20, 16] has to be used to achieve robust information. However, the aim of this paper is to define and investigate an appropriate local similarity function which makes use of structural and geometric information and to derive ....

J. Aloimonos and D. Shulman. Integration of Visual Modules --- An extension of the Marr Paradigm. Academic Press, London, 1989.

....at the results from biology to become inspired for suitable de nitions of constraints. Another important problem the visual system has to cope with is an extremely high degree of vagueness and uncertainty in its low level processes such as edge detection, optic ow analysis and stereo estimation [1]. However, by integrating information across visual modalities (see, e.g. 32, 8] the human visual systems acquires visual representations which allows for actions with high precision and certainty within the 3D world even under rather uncontrolled conditions . The power of modality fusion ....

J. Aloimonos and D. Shulman. Integration of Visual Modules | An extension of the Marr Paradigm. Academic Press, London, 1989.

....facilitating the implementation of large integrated systems. Another important issue to be considered in integration is that of the emergent behavior of the system. The vision modules often can not individually solve their assigned tasks unless they are given a specific model of the reality [1]. When one considers such an implementation of a single vision module, each simplifying assumption re1 quired by the module drastically reduces the scope of its application to real images. However, in an integrated system, the availability of additional cues lets the complete system transcend ....

Y. Aloimonos and D. Shulman. Integration of Visual Modules: An Extension of the Marr Paradigm. Academic Press, San Diego, CA, 1989.

.... or motion are evaluated independently in almost completely separate pathways [21, 5, 10] Curvature is supposed to be processed by such a separate channel in early vision, too [23, 2, 3] Methods to reconstruct a three dimensional scene from single features are known as shape from X methods [6, 8, 1, 9]. Usually, they lead to so called illposed problems since the mapping of a three dimensional image onto a gray scale (contour, texture, image has no unique inversion. A well known approach to deal with this problem is called regularization. It removes the non uniqueness by imposing a ....

J. Aloimonos and D. Shulman. Integration of Visual Modules --- An Extension of the Marr Paradigm. Academic Press, San Diego London, 1989.

....machine learning community have looked at both classification and symbolic inference, but not at the combination of the two. Syntactic Approach to Integration In terms of the integration of visual modules, SLS is on the opposite end of a spectrum from the approach taken by Aloimonos and Shulman [1]. Aloimonos and Shulman recommend analyzing the semantics of each visual operator closely, and integrating them by merging their mathematical constraints. Although such an approach is very powerful in theory, it is difficult, if not impossible, in practice. SLS, on the other hand, takes an ....

J. Aloimonos and D. Shulman. Integration of Visual Modules: An Extension of the Marr Paradigm. Academic Press, Inc., Boston, 1989.

.... with the photometric and geometric issues of image formation, psychologists, biologists and neuroscientists [Edelman, 1989; Gregory, 1970; Kanizsa, 1979] studied how biological organisms perform visual perception, and engineers, mathematicians and computer scientists [Ballard and Brown, 1982; Aloimonos and Shulman, 1989; Marr, 1982] investigated the development and computation of representations of the visual environment such as shape, form and color. Since all biological and robot systems exist in space time and move in their environments, in order to capture all dimensions of objective reality, they must ....

....main problem with all these smoothing techniques is the existence of motion boundaries. This realization has led to various studies dealing explicitly with discontinuities, which follow one of two general approaches. Discontinuous regularization techniques [Shulman and Herv e, 1989; Shulman, 1990; Aloimonos and Shulman, 1989] reconstruct the optical flow by smoothing as little as possible near discontinuities. Probabilistic approaches such as Markov random field techniques [Geman and Geman, 1984] localize the discontinuities, or more exactly the places where the probability of having a discontinuity is highest, in ....

[Article contains additional citation context not shown here]

J. Aloimonos and D. Shulman. Integration of Visual Modules: An Extension of the Marr Paradigm. Academic Press, Boston, 1989.

.... Israel Institute of Technology Technion Haifa 32000, Israel 2 Laboratory of Image Analysis, Aalborg University Aalborg East, Denmark Abstract The visual robustness of biological systems is in part due to their ability to actively integrate (fuse) information from a number of visual cues [2], 29] In addition to active integration, the perception action nature of biological vision demands event driven behavioral composition. Providing mechanical vision systems with similar capabilities therefore requires tools and techniques for cue integration and behavioral composition. In this ....

.... [5] 6] In addition to adjusting parameters, This work was supported in part by the European Community Israel collaboration ECIS 003 project 2 the visual robustness of biological systems is also due to their ability to actively integrate (fuse) 1 information from a number of visual cues [2], 29] as well as to compose behaviors in response to dynamic events. Providing mechanical active vision systems with similar capabilities therefore requires tools and techniques for cue integration and behavioral composition. In this paper, we focus on two important features of a robust vision ....

[Article contains additional citation context not shown here]

Y. Aloimonos and D. Shulman. Integration of Visual Modules: An Extension of the Marr Paradigm. Academic Press, 1989.

.... University Aalborg East, Denmark fpaolo,hicg vision.auc.dk Abstract It is becoming increasingly clear that the robustness demonstrated by biological vision systems is in large part due to their ability to actively integrate (fuse) 1 information from a number of similar or differing visual cues [34, 2]. In addition to active integration, the perception action nature of biological systems relies upon the ability to effectively compose behaviors in an event driven manner. Providing mechanical vision systems with capabilities similar to those possessed by biological systems therefore entails ....

.... it is becoming increasingly clear that a robust vision system should make use of a seamless integration of a number of functionally equivalent (homogeneous) or nonequivalent (non homogeneous) modules [34] Indeed, integration forms the basis of a recently proposed extension of the Marr paradigm [2] in which the authors stress that future progress in computer vision will be a result of module integration: Now that most of the modules have been studied in isolation, we think that it is time that they be tested in pairs, triples, and so on . The advantages of integration are evident in ....

[Article contains additional citation context not shown here]

Y. Aloimonos and D. Shulman. Integration of Visual Modules: An Extension of the Marr Paradigm. Academic Press, 1989.

No context found.

J. Aloimonos and D. Shulman. Integration of Visual Modules: An Extension of the Marr Paradigm. Academic Press, Boston, 1989.

....as affine, quadratic, high level polynomial and the like. Some popular approaches are based on smoothness, i.e. no parametric model is assumed but instead the assumption is made that the flow field in the patch is a smooth function and the solution is obtained through a regularization procedure [1]. The problem with all these approaches is that they work adequately when the image patch considered is the image of a smooth scene patch, but they lead to wrong estimates when the scene patch contains depth discontinuities, because the values of the flow on the two sides of the discontinuity are ....

J. Aloimonos and D. Shulman. Integration of Visual Modules: An Extension of the Marr Paradigm. Academic Press, Boston, 1993.

....judgements and actions (segmentation, recognition, manipulation) Punctate reconstruction ( Physics based vision ) has been for some time the dominant paradigm in computer vision. Reconstructionism does not rule out active techniques, and the work is growing ever more sophisticated (e.g. [Aloimonos and Schulman, 1989] ) Reconstructionism usually comes with a bottom up, datadriven approach to general vision. This is good: high level knowledge (or wishful thinking, or probability) has nothing to do with low level vision. Methodologically, however, reconstructionismcan lead to unjustified assumptions of linearly ....

J. Aloimonos and D. Schulman. Integration of Visual Modules: An Extension of the Marr Paradigm. Academic Press, Boston, 1989.

No context found.

Y. Aloimonos and D. Shulman, Integration of Visual Modules --- An extension of the Marr Paradigm, Academic Press, London, 1989.

No context found.

Y. Aloimonos and D. Shulman. Integration of Visual Modules | An extension of the Marr Paradigm. Academic Press, London, 1989.

No context found.

Y. Aloimonos and D. Shulman, Integration of Visual Modules --- An extension of the Marr Paradigm, Academic Press, London, 1989.

No context found.

Aloimonos, J., and Shulman, D. Integration of Visual Modules An Extension of the Marr Paradigm. Academic Press, 1989.

No context found.

J. Aloimonos and David Shulman, Integration of Visual Modules An Extension of the Marr Paradigm. Academic Press, 1989.

No context found.

Y. Aloimonos and D. Shulman. Integration of Visual Modules: An Extension of the Marr Paradigm. Academic Press, 1989.

No context found.

Aloimonos, J., and Shulman, D. Integration of Visual Modules An Extension of the Marr Paradigm. Academic Press, 1989.

No context found.

J. Aloimonos and David Shulman, Integration of Visual Modules An Extension of the Marr Paradigm. Academic Press, 1989.

No context found.

J. Aloimonos and D. A. Shulman, Integration of Visual Modules: an Extension of the Marr Paradigm (Academic, New York, 1989).

No context found.

Aloimonos, J. 1989. Integration of visual modules : an extension of the Marr paradigm. Boston, Mass: Academic Press.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC