Steven Lehar - Academia.edu (original) (raw)
Papers by Steven Lehar
Perception, Apr 1, 2003
Introduction Visual illusions and perceptual grouping phenomena offer an invaluable tool for prob... more Introduction Visual illusions and perceptual grouping phenomena offer an invaluable tool for probing the computational mechanism of low-level visual processing. Some illusions, like the Kanizsa figure shown in figure 1a, reveal a process of collinear illusory-contour formation, whereby the illusory contour spans the gap between stimulus edges that are both parallel, and spatially aligned. This kind of illusory-contour formation has been modeled by neural network models that make use of`cooperative cells' equipped with elongated spatial receptive fields designed to detect and, by top^down feedback, to complete the collinear alignment with a line of neural activation along the perceived illusory contour (Grossberg and Mingolla 1985; Zucker et al 1989). There are, however, other illusory groupings which are not so easy to account for in neutral network terms. For example, the Ehrenstein illusion, shown in figure 1b, exhibits an illusory contour orthogonal to the stimulus line segments instead of collinear with them. Grossberg and Mingolla (1985) account for this kind of illusory contour with a competitive interaction between cooperative cells at any particular spatial location whose orientations differ
The Kluwer international series in engineering and computer science, 2000
A serious crisis is identified in theories of neurocomputation, which has implications for comput... more A serious crisis is identified in theories of neurocomputation, which has implications for computer models of visual processing inspired by biological vision. The problem is reflected in a persistent disparity between the phenomenological or experiential account of visual perception and the neurophysiological or computational level of description. In particular, conventional concepts of neural processing as well as image processing algorithms offer no explanation for the holistic global aspects of perception identified by Gestalt theory. The problem is paradigmatic, and can be traced to atomistic concepts of computation embodied in the biological notion of neurocomputation, as well as in the paradigm of digital computation. I propose a perceptual modeling approach, i.e. to model the percept as experienced subjectively, rather than the objective neurophysiological state of the visual system that supposedly subserves that experience. A Gestalt Bubble model is presented to demonstrate how the elusive Gestalt principles of emergence, reification, and invariance, can be expressed in a quantitative model of the subjective experience of visual consciousness. That model in turn reveals a unique computational strategy underlying visual processing, which is unlike any algorithm devised by man, and certainly unlike the atomistic feed-forward model of visual processing characteristic of many biological and computer models. The perceptual modeling approach reveals the primary function of perception as that of generating a fully spatial virtual-reality replica of the external world in an internal representation.
ABSTRACT An abstract is not available.
Certain recent neural models of natural vision systems are defined in sufficiently concrete terms... more Certain recent neural models of natural vision systems are defined in sufficiently concrete terms as to be immediately applicable to image processing tasks. In our work, the Boundary Contour System (BCS) and Feature Contour System (FCS) human vision models developed at the Center for Adaptive Systems by Grossberg and Mingolla, and Grossberg and Todorovic are applied to magnetic resonance brain scan imagery to highlight significant features for the purpose of pattern recognition.' 1. We would like to thank Steve Grossbeg and Ennio Mingolla for their help in this work.
Abstract One of the enduring mysteries of vision is the question whether the world we see in visu... more Abstract One of the enduring mysteries of vision is the question whether the world we see in visual consciousness is the real world itself, or whether it is merely a perceptual replica of that world in an internal representation: The theory of Direct Perception as opposed to Indirect Perception. Direct perception, also known as Naive Realism, is the natural intuitive understanding of vision that we assume from infanthood that the world we see around us is the world itself. The physiology of the visual system, however, from the eye, through the optic nerve to the brain, suggests an indirect or representational principle of perception, whereby the sensory information flows inward from the world through the eye to the brain. There is no way that the brain can possibly be aware of objects in the external world directly, our experience is necessarily confined to internal states of the brain as they respond to sensory input. Other theories have been proposed to account for the external nature of experience, from projection theories, whereby experience is projected out of the brain to appear back in the world, to eliminative theories whereby experience is a purely subjective nonphysical entity that has no location either within the brain or beyond it. These theories are unfalsifiable because the experience that is either projected or exists in some orthogonal space cannot be detected by physical means because their prediction that experience as a nonphysical entity, cannot be detected, is identical to the null prediction that the nonphysical aspect of experience does not exist. If we are to accept a materialist explanation of mind as a physical process taking place within the physical mechanism of the brain, then experience must be a physical phenomenon, detectable in principle by physical means, and it cannot be located anywhere else but in the physical brain that generates and sustains it.
Journal of Vision, Mar 14, 2010
SPIE Proceedings, Aug 1, 1991
ABSTRACT Artificial vision algorithms suffer from the multiple ambiguities present in natural ima... more ABSTRACT Artificial vision algorithms suffer from the multiple ambiguities present in natural imagery. Biological vision exhibits a robust adaptability in the presence of such ambiguities. The Multiple Resonant Boundary Contour System (MRBCS) is animage processing algorithm modeled on the Boundary Contour System1 (BCS), a neural model of natural vision which has been confirmed by its ability to explain and predict numerous psychophysical illusions. Interesting properties of natural visual processing mechanisms are revealed through simulation using natural imagery. 2. INTRODUCTION Attempts to implement artificial vision are plagued by the multiple ambiguities inherent in natural imagery. There are extraneous pieces of information that must be removed, and missing pieces that must be restored, in order to provide a reliablegeometrical representation ready for robust recognition algorithms. Each ambiguity can be handled using discrete processingmodules, and the image processing literature provides a wide range of techniques to achieve this end. For instance: image
Springer eBooks, 1990
Recent neural models of natural vision systems are defined in sufficiently concrete terms as to b... more Recent neural models of natural vision systems are defined in sufficiently concrete terms as to be immediately applicable to practical image processing tasks. In particular the Boundary Contour System and Feature Contour System human vision models developed by Grossberg and Mingolla, and Grossberg and Todorovicz are applied to satellite weather imagery to highlight significant features for the purpose of pattern recognition.
Contents: Preface. The Two Worlds of Reality. The Dimensions of Conscious Experience. The Enigma ... more Contents: Preface. The Two Worlds of Reality. The Dimensions of Conscious Experience. The Enigma of Gestalt Phenomena. The Computational Mechanism of Perception. The Perception of Illumination. Recognition vs. Completion, Abstraction vs. Reification. Relation to Neurophysiology. Harmonic Resonance Theory. Image Theory of Language and Cognition. Motor Control and Field Theory. A Psycho-Aesthetic Hypothesis. Appendices: The Coplanarity Field. The Occlusion Field. The Orthogonality Field. Edge Consistency and Inconsistency Constraints. Influence of the Visual Input.
IEEE Transactions on Knowledge and Data Engineering, Apr 1, 1992
Behavioral and Brain Sciences, Aug 1, 2003
Inji gwahag jag'eob, Jun 1, 2018
There are four prominent properties of the mind that pose the greatest challenge to neuroscience ... more There are four prominent properties of the mind that pose the greatest challenge to neuroscience and the mind-brain problem. The first is the unity of conscious experience: We experience every object in perception at a specific location in the global sphere of surrounding experience, and the whole assembly of perceptions hangs together as a single unified structure. The second great mystery is the manifestly pictorial nature of visual experience: We see the world as a surrounding structure that is explicitly three-dimensional and spatial. The third great mystery is the holistic nature of perception as revealed by Gestalt theory, or the way that the global percept emerges from the parallel influence of countless individual features simultaneously. The fourth is the invariance evident in perception, whereby objects maintain their structural integrity and recognized identity even as they rotate, translate, and scale by perspective in their motions through the world. The world itself appears stable, even as our head and eyes and brain rotate relative to that world. These four perplexing properties of mind pose such a profound challenge to theories of brain function as a basis for mind, that historically they have been largely ignored, if not actively "explained away", as if they had no relevance to the mind-brain problem. I propose that these four mysterious properties of mind are not unrelated, but in fact they are intimately related, and they collectively implicate a unified, pictorial, holistic, and invariant principle of computation and representation in the brain. Far from ignoring these most perplexing properties of mind, neurscience would do well to pay close attention to the computational principles that they implicate collectively.
Psychology Press eBooks, Nov 23, 2020
Behavioral and Brain Sciences, Aug 1, 2003
To avoid endless and futile debate, critics of an alternative paradigmatic hypothesis cannot simp... more To avoid endless and futile debate, critics of an alternative paradigmatic hypothesis cannot simply state their own paradigmatic assumptions as if they were plain fact while dismissing those of the opposition as self-evidently absurd, because it is exactly those initial assumptions that are brought into question by the paradigmatic proposal. Perceived incredibility is no valid ground for rejection of a paradigm whose alternatives are at least equally incredible, and arguably more so.
Consciousness and Cognition, 2000
Psychology was originally formulated as the science of the psyche, i.e. the subjective side of th... more Psychology was originally formulated as the science of the psyche, i.e. the subjective side of the mind / brain barrier. However time and again it has been diverted from this objective in the supposed interest of scientific rigor. The Behaviorists proposed to transform psychology to a science of behavior, and today the Neuroreductionists propose to transform it to a science of neurophysiology. In the process they attempt to deny the very existence of conscious experience as valid object of scientific scrutiny. However the subjective conscious experience is a primary source of evidence for the nature of the representation in the brain. I propose a quantitative phenomenolgy to express the dimensions of conscious experience in information theoretic terms. This approach leads to interesting observations of the properties of phenomenal perspective, that clearly reveal the phenomenal world as an internal rather than external entity.
Perception, Apr 1, 2003
Introduction Visual illusions and perceptual grouping phenomena offer an invaluable tool for prob... more Introduction Visual illusions and perceptual grouping phenomena offer an invaluable tool for probing the computational mechanism of low-level visual processing. Some illusions, like the Kanizsa figure shown in figure 1a, reveal a process of collinear illusory-contour formation, whereby the illusory contour spans the gap between stimulus edges that are both parallel, and spatially aligned. This kind of illusory-contour formation has been modeled by neural network models that make use of`cooperative cells' equipped with elongated spatial receptive fields designed to detect and, by top^down feedback, to complete the collinear alignment with a line of neural activation along the perceived illusory contour (Grossberg and Mingolla 1985; Zucker et al 1989). There are, however, other illusory groupings which are not so easy to account for in neutral network terms. For example, the Ehrenstein illusion, shown in figure 1b, exhibits an illusory contour orthogonal to the stimulus line segments instead of collinear with them. Grossberg and Mingolla (1985) account for this kind of illusory contour with a competitive interaction between cooperative cells at any particular spatial location whose orientations differ
The Kluwer international series in engineering and computer science, 2000
A serious crisis is identified in theories of neurocomputation, which has implications for comput... more A serious crisis is identified in theories of neurocomputation, which has implications for computer models of visual processing inspired by biological vision. The problem is reflected in a persistent disparity between the phenomenological or experiential account of visual perception and the neurophysiological or computational level of description. In particular, conventional concepts of neural processing as well as image processing algorithms offer no explanation for the holistic global aspects of perception identified by Gestalt theory. The problem is paradigmatic, and can be traced to atomistic concepts of computation embodied in the biological notion of neurocomputation, as well as in the paradigm of digital computation. I propose a perceptual modeling approach, i.e. to model the percept as experienced subjectively, rather than the objective neurophysiological state of the visual system that supposedly subserves that experience. A Gestalt Bubble model is presented to demonstrate how the elusive Gestalt principles of emergence, reification, and invariance, can be expressed in a quantitative model of the subjective experience of visual consciousness. That model in turn reveals a unique computational strategy underlying visual processing, which is unlike any algorithm devised by man, and certainly unlike the atomistic feed-forward model of visual processing characteristic of many biological and computer models. The perceptual modeling approach reveals the primary function of perception as that of generating a fully spatial virtual-reality replica of the external world in an internal representation.
ABSTRACT An abstract is not available.
Certain recent neural models of natural vision systems are defined in sufficiently concrete terms... more Certain recent neural models of natural vision systems are defined in sufficiently concrete terms as to be immediately applicable to image processing tasks. In our work, the Boundary Contour System (BCS) and Feature Contour System (FCS) human vision models developed at the Center for Adaptive Systems by Grossberg and Mingolla, and Grossberg and Todorovic are applied to magnetic resonance brain scan imagery to highlight significant features for the purpose of pattern recognition.' 1. We would like to thank Steve Grossbeg and Ennio Mingolla for their help in this work.
Abstract One of the enduring mysteries of vision is the question whether the world we see in visu... more Abstract One of the enduring mysteries of vision is the question whether the world we see in visual consciousness is the real world itself, or whether it is merely a perceptual replica of that world in an internal representation: The theory of Direct Perception as opposed to Indirect Perception. Direct perception, also known as Naive Realism, is the natural intuitive understanding of vision that we assume from infanthood that the world we see around us is the world itself. The physiology of the visual system, however, from the eye, through the optic nerve to the brain, suggests an indirect or representational principle of perception, whereby the sensory information flows inward from the world through the eye to the brain. There is no way that the brain can possibly be aware of objects in the external world directly, our experience is necessarily confined to internal states of the brain as they respond to sensory input. Other theories have been proposed to account for the external nature of experience, from projection theories, whereby experience is projected out of the brain to appear back in the world, to eliminative theories whereby experience is a purely subjective nonphysical entity that has no location either within the brain or beyond it. These theories are unfalsifiable because the experience that is either projected or exists in some orthogonal space cannot be detected by physical means because their prediction that experience as a nonphysical entity, cannot be detected, is identical to the null prediction that the nonphysical aspect of experience does not exist. If we are to accept a materialist explanation of mind as a physical process taking place within the physical mechanism of the brain, then experience must be a physical phenomenon, detectable in principle by physical means, and it cannot be located anywhere else but in the physical brain that generates and sustains it.
Journal of Vision, Mar 14, 2010
SPIE Proceedings, Aug 1, 1991
ABSTRACT Artificial vision algorithms suffer from the multiple ambiguities present in natural ima... more ABSTRACT Artificial vision algorithms suffer from the multiple ambiguities present in natural imagery. Biological vision exhibits a robust adaptability in the presence of such ambiguities. The Multiple Resonant Boundary Contour System (MRBCS) is animage processing algorithm modeled on the Boundary Contour System1 (BCS), a neural model of natural vision which has been confirmed by its ability to explain and predict numerous psychophysical illusions. Interesting properties of natural visual processing mechanisms are revealed through simulation using natural imagery. 2. INTRODUCTION Attempts to implement artificial vision are plagued by the multiple ambiguities inherent in natural imagery. There are extraneous pieces of information that must be removed, and missing pieces that must be restored, in order to provide a reliablegeometrical representation ready for robust recognition algorithms. Each ambiguity can be handled using discrete processingmodules, and the image processing literature provides a wide range of techniques to achieve this end. For instance: image
Springer eBooks, 1990
Recent neural models of natural vision systems are defined in sufficiently concrete terms as to b... more Recent neural models of natural vision systems are defined in sufficiently concrete terms as to be immediately applicable to practical image processing tasks. In particular the Boundary Contour System and Feature Contour System human vision models developed by Grossberg and Mingolla, and Grossberg and Todorovicz are applied to satellite weather imagery to highlight significant features for the purpose of pattern recognition.
Contents: Preface. The Two Worlds of Reality. The Dimensions of Conscious Experience. The Enigma ... more Contents: Preface. The Two Worlds of Reality. The Dimensions of Conscious Experience. The Enigma of Gestalt Phenomena. The Computational Mechanism of Perception. The Perception of Illumination. Recognition vs. Completion, Abstraction vs. Reification. Relation to Neurophysiology. Harmonic Resonance Theory. Image Theory of Language and Cognition. Motor Control and Field Theory. A Psycho-Aesthetic Hypothesis. Appendices: The Coplanarity Field. The Occlusion Field. The Orthogonality Field. Edge Consistency and Inconsistency Constraints. Influence of the Visual Input.
IEEE Transactions on Knowledge and Data Engineering, Apr 1, 1992
Behavioral and Brain Sciences, Aug 1, 2003
Inji gwahag jag'eob, Jun 1, 2018
There are four prominent properties of the mind that pose the greatest challenge to neuroscience ... more There are four prominent properties of the mind that pose the greatest challenge to neuroscience and the mind-brain problem. The first is the unity of conscious experience: We experience every object in perception at a specific location in the global sphere of surrounding experience, and the whole assembly of perceptions hangs together as a single unified structure. The second great mystery is the manifestly pictorial nature of visual experience: We see the world as a surrounding structure that is explicitly three-dimensional and spatial. The third great mystery is the holistic nature of perception as revealed by Gestalt theory, or the way that the global percept emerges from the parallel influence of countless individual features simultaneously. The fourth is the invariance evident in perception, whereby objects maintain their structural integrity and recognized identity even as they rotate, translate, and scale by perspective in their motions through the world. The world itself appears stable, even as our head and eyes and brain rotate relative to that world. These four perplexing properties of mind pose such a profound challenge to theories of brain function as a basis for mind, that historically they have been largely ignored, if not actively "explained away", as if they had no relevance to the mind-brain problem. I propose that these four mysterious properties of mind are not unrelated, but in fact they are intimately related, and they collectively implicate a unified, pictorial, holistic, and invariant principle of computation and representation in the brain. Far from ignoring these most perplexing properties of mind, neurscience would do well to pay close attention to the computational principles that they implicate collectively.
Psychology Press eBooks, Nov 23, 2020
Behavioral and Brain Sciences, Aug 1, 2003
To avoid endless and futile debate, critics of an alternative paradigmatic hypothesis cannot simp... more To avoid endless and futile debate, critics of an alternative paradigmatic hypothesis cannot simply state their own paradigmatic assumptions as if they were plain fact while dismissing those of the opposition as self-evidently absurd, because it is exactly those initial assumptions that are brought into question by the paradigmatic proposal. Perceived incredibility is no valid ground for rejection of a paradigm whose alternatives are at least equally incredible, and arguably more so.
Consciousness and Cognition, 2000
Psychology was originally formulated as the science of the psyche, i.e. the subjective side of th... more Psychology was originally formulated as the science of the psyche, i.e. the subjective side of the mind / brain barrier. However time and again it has been diverted from this objective in the supposed interest of scientific rigor. The Behaviorists proposed to transform psychology to a science of behavior, and today the Neuroreductionists propose to transform it to a science of neurophysiology. In the process they attempt to deny the very existence of conscious experience as valid object of scientific scrutiny. However the subjective conscious experience is a primary source of evidence for the nature of the representation in the brain. I propose a quantitative phenomenolgy to express the dimensions of conscious experience in information theoretic terms. This approach leads to interesting observations of the properties of phenomenal perspective, that clearly reveal the phenomenal world as an internal rather than external entity.