Marr's computational approach to vision

  title={Marr's computational approach to vision},
  author={Tomaso A. Poggio},
  journal={Trends in Neurosciences},
  • T. Poggio
  • Published 31 December 1981
  • Art
  • Trends in Neurosciences

Categorization through temporal analysis of patterns

This book is considered by many in the field of computer vision as the main work of these last fifteen years and was the first to propose a complete methodology for computer vision which became known as the Marr paradigm.

Recent advances in the understanding of vision

This review surveys a large part of the broad subject of ‘How the authors see’ from several levels of understanding: component and circuit analysis of the system; mechanisms and algorithms of the visual process; and computational theory of vision.

The Levels of Understanding Framework, Revised

The “levels of understanding” framework described in Marr's Vision is discussed and an updated version to capture the changes in computation and neuroscience over the last 30 years is proposed.

Neural computation of depth from binocular disparity

It is found that disparity encoding in primary visual cortex can be explained by shallow, feed-forward networks optimized to extract absolute depth from naturalistic images, and evidence for a differential contribution of cortical layers in stereoscopic depth perception is reported.

Deep effect trajectory prediction in robot manipulation

A Robust and Discriminative Feature Representation based on Compact Coding

This work proposes a feature combining framework, which benefits from the recent fruitful research on compact coding of visual features, and combines colour histogram and texture under the framework, and experiments show that the method provides an effective way to the representation of features and has a wide application.

Marr’s Computational Level and Delineating Phenomena

A key component of scientific inquiry, especially inquiry devoted to developing mechanistic explanations, is delineating the phenomenon to be explained. The task of delineating phenomena, however,

3D Manipulation of Objects in Photographs

This thesis describes a system that allows users to to perform full three-dimensional manipulations to objects in photographs. Cameras and photo-editing tools have contributed to the explosion in

Applications of the Relative Neighbourhood Graph

This paper provides a review of some of the applications of the relative neighborhood graph, including wireless network communications, archaeological network analysis, grid typification in cartography, data mining for geographic information systems, shape analysis, image morphology, polygon decomposition, the extraction of primal sketches in computer vision, and more.

Perceptual Representation Content, Constancies and Phenomenal Consciousness

ion‖ (Burge, 2010, p. 381). 20 Specifically, Burge talks about the singular and attributive elements in perception. Perceptual representational contents have singular and attributive elements. I will



The Interpretation of Visual Motion

This book uses the methodology of artificial intelligence to investigate the phenomena of visual motion perception: how the visual system constructs descriptions of the environment in terms of

Cooperative computation of stereo disparity.

It is shown that this algorithm successfully extracts information from random-dot stereograms, and its implications for the psychophysics and neurophysiology of the visual system are briefly discussed.

Smallest channel in early human vision.

We suggest from the psychophysical data on two-point and line acuity that the smallest foveal channel in human vision must have an excitatory center with a diameter of around 1' 20". Taking into

Visual Information Processing: The Structure and Creation of Visual Representations

An overall framework is suggested for extracting shape information from images, in which the analysis proceeds through three representations: the primal sketch, which makes explicit the intensity changes and local two-dimensional geometry of an image, the 2 1/2-D sketch, and the 3-D model representation, which allows an object-centred description of the three-dimensional structure and organization of a viewed shape.

Bandpass channels, zero-crossings, and early visual information processing.

Under appropriate conditions zero-crossings of a bandpass signal are very rich in information and may contain sufficient information for much of the subsequent processing in the early stages of visual information processing.

Information in the zero crossings of bandpass signals

  • B. Logan
  • Mathematics
    The Bell System Technical Journal
  • 1977
An interesting subclass of bandpass signals {h} is described wherein the zero crossings of h determine h within a multiplicative constant. The members may have complex zeros, but it is necessary that

Thinking About Seeing

The representation of three-dimensional objects

Three representations are explored which have varying time and space tradeoffs and one of the representations simplifies the storage of dynamic objects and objects with redundant sub-parts.