J. Tanner and C.A. Mead, "An integrated analog optical motion sensor," in VLSI Signal Processing, 2, S.-Y. Kung, R. Owen, and G. Nash, Eds., p. 59 ff. IEEE Press, 1986.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....known to be robust in the face of noise and aliasing, they give reliable measurements of velocity and they allow an easy treatment of the aperture problem. However, due to the complexity of implementing a battery of spatiotemporal filters, current VLSI motion chips use gradient based algorithms [24], 6] or correlation based algorithms [12] 7] 9] as they can be implemented within very compact circuits. In this paper we present a new energy based algorithm that significantly minimizes the complexity of these kinds of methods. The reduction of complexity is due to: 1) the use of a new ....

J. Tanner and C. Mead. An integrated analog optical motion sensor. In R.W. Brodersen and H.S. Moscovitz, editors, VLSI Signal Processing,volume 2, pages 59--87, New York, 1988. IEEE.

....are known to be robust to noise and aliasing, they give reliable measurements of velocity and they allow an easy treatment of the 41 aperture problem. However, due to the complexity of implementing a battery of spatiotemporal filters, current VLSI motion chips use gradient based algorithms [82, 24] or correlation based algorithms [46, 25, 33] as they can be implemented within very compact circuits. In this chapter we present energy based algorithm that significantly minimizes the complexity of these kinds of methods. The reduction of complexity is due to: 1) the use of a new wideband ....

....is computed as the shift that yields the best fit between image regions at different times. ffl Gradient based: They compute velocity from spatiotemporal derivatives of image intensity [37] Implementations on VLSI yield to circuits with high sensibility to noise and spatiotemporal aliasing [24, 82]. ffl Energy based: They compute velocity by combination of the energy outputs of a set of spatiotemporal filters tuned to several velocities or spatiotemporal frequencies [30, 31] ffl Correlation based: velocity is obtained by combining the output of motion sensors based on the correlation of ....

J. Tanner and C. Mead. An integrated analog optical motion sensor. In R.W. Brodersen and H.S. Moscovitz, editors, VLSI Signal Processing, volume 2, pages 59-87, New York, 1988.

....edge triggers them (figure 2b) The spike is used to generate a digital pulse in each arm of the detector pair. The digital pulses on sets are measured in time to com Architecture Functionality Transistors per pixel (incl. I O) Pixel size m 2 Optical Motion Sensor Tanner and Mead, 1989, [2] Gradient computation of optical flow using a resistive grid 71 180 x 180 32,400 Correlation based, velocity tuned pixels motion sensor Delbrck, 1993, 3] Tuned to a single spatio temporal frequency 41 135 x 135 18,225 Temporal Domain Optical Flow Sensor Ettiene Cummings et al., ....

J. Tanner and C. Mead, "An Integrated Analog Optical Motion Sensor," in VLSI Signal Processing II, S.Y. Kung, Ed. IEEE Press, 1986, pp. 59-76.

.... , Parsytec transputer [42] and hybrid pyramidal vision machine (AIS4000 and CSA transputer) 13] high cost, weight and power consumption Special image processing hardware PIPE [1] 11] 39] 46] Datacube [34] and PRISM 3 [36] low precision Dedicated VLSI chips Vision Chips: gradient method [31] [44] , correspondence method [12] 43] and biological receptive field design [14] 30] low resolution Non Vision Chips: analog neural networks [20] digital block matching technique [4] 18] 49] coarsely quantized estimates Table 3: Real time motion estimation algorithms algorithmic approach ....

....there is significant potential in this approach because of its small size and power consumption. One method uses analog computing circuitry with built in photoreceptors. These estimation chips represent a major class in the so called Vision Chips or Seeing Silicons [29] Gradient methods [31] [44] , correspondence methods [12] 43] and biological receptive field designs [14] 30] have all been implemented. Although a great deal of progress has been made in the past decade, there are still several barriers to this approach. First, there is a lack of VLSI friendly 6 algorithms. Most flow ....

Tanner, J. and Mead, C., "An Integrated Analog Optical Motion Sensor", VLSI Signal Processing II, R.W. Brodersen and H.S. Moscovitz, Eds., pp. 59-87, New York, 1988.

....of providing 2D direction and or velocity. The sensors in our survey are summarized in Table I. An extensive survey of motion sensors, including 1D sensors, can be found in [3] Any review of continuous time motion sensors must begin with Tanner and Mead s 1986 gradient based optical flow chip [4], which provided an estimate of the global 2D velocity by satisfying a global image constancy constraint in an 8 Theta8 photoreceptor array. The chip worked for high contrasts over a small velocity range, and suffered from fixed pattern noise due to low precision photoreceptors and derivative ....

....a high fill factor photoreceptor array, an edge detector array, and a motion processing array. Due to the wiring required between the three stages of processing, this design cannot be HIGGINS, DEUTSCHMANN AND KOCH: PULSE BASED 2D MOTION SENSORS 2 Authors Reference Year Function Tanner and Mead [4] 1986 Global velocity Delbruck [5] 1993 Semi local direction of motion Etienne Cummings et al. 6] 1997 Local velocity Deutschmann et al. 8] 1998 Local direction of motion TABLE I Review of integrated 2D motion sensors easily scaled to large array sizes. The largest fabricated array of these ....

J. Tanner and C. Mead, "An integrated analog optical motion sensor," in VLSI Signal Processing, II, New York, Ed., pp. 59-- 76. IEEE Press, 1986.

.... [39] Parsytec transputer [32] and hybrid pyramidal vision machine (AIS 4000 and CSA transputer) 11] high cost, weight and power consumption Image processing hardware PIPE [1] 9] 29] 35] Datacube [25] and PRISM 3 [27] low precision Dedicated VLSI chips Vision Chips: gradient method [23] [34] , correspondence method [10] 33] and biological receptive field design [12] 22] low resolution Non Vision Chips: analog neural networks [17] digital block matching technique [3] 15] 38] coarsely quantized estimates The hardware approach uses specialized hardware to achieve real time ....

Tanner, J. and Mead, C., "An Integrated Analog Optical Motion Sensor", VLSI Signal Processing II, R.W. Brodersen and H.S. Moscovitz, Eds., pp. 59-87, New York, 1988.

.... Analog Network Alan Stocker Institute of Neuroinformatics University and ETH Zurich Winterthurerstrasse 190 8057 Zurich, Switzerland Rodney Douglas Institute of Neuroinformatics University and ETH Zurich Winterthurerstrasse 190 8057 Zurich, Switzerland Abstract In 1986, Tanner and Mead [1] implemented an interesting constraint satisfaction circuit for global motion sensing in aVLSI. We report here a new and improved aVLSI implementation that provides smooth optical flow as well as global motion in a two dimensional visual field. The computation of optical flow is an ill posed ....

J. Tanner and C.A. Mead. An integrated analog optical motion sensor. In S.-Y. Kung, R. Owen, and G. Nash, editors, VLSI Signal Processing, 2, page 59 ff. IEEE Press, 1986.

....extreme solutions. One possibility was mentioned in section 3.2, but this has to be carefully studied in relation with the requirements of a particular application. PI n1220 22 Etienne M emin and Tanguy Risset The idea of integrating a processor behind a sensor has been studied for a while [27], and the concept of smart sensors [7] and Active Pixel Sensor [20] led to numerous achievements. Bi dimensional arrays of such sensors have already been implemented [1] However, because of current technology restriction, these smart sensors are often implemented with full custom design (often in ....

J. Tanner and C. Mead. An integrated analog optical motion sensor. In S.Y. Kung, R. Owen, and J.G. Nash, editors, VLSI signal processing II. IEEE Press, 1986.

....motion sensors that incorporate the photoreceptors and the processing circuitry on a single chip: Some early implementations used intensity based methods. Among the first circuits operating on a gradient algorithm was an optical mouse chip for estimating an uniform velocity in two dimensions [10] 1 . Under an orthogonal projection system, uniform velocity corresponds to global translation of a rigid object space relative to the sensor. The circuit consisted of an 8 Theta8 array of velocity sensors. Offset effects of individual pixels were reduced by averaging across the array. The ....

....[20, 23] But spatial edge detectors tend to consume more silicon area and exhibit more offset problems, lower sensitivity, and slower response than the temporal edge detector described here. Most of the previously reported analog VLSI motion sensors only respond robustly to high contrast stimuli [10, 19], show a contrast dependent output [14, 15, 18, 22, 25] or have a non monotonic response curve [14, 15, 18, 19, 22, 25] A few implementations were reported to unambiguously encode velocity down to low contrasts [17, 23, 24, 26] but only for [26] results are given in the reference. How can future ....

J. Tanner and C. Mead, "An Integrated Analog Optical Motion Sensor," in VLSI Signal Processing, II, S. Y. Kung, R. E. Owen and J. G. Nash, Eds. New York: IEEE Press, 1986, pp. 59-76.

....parallel image processing architectures implemented in analog VLSI technology. Within this framework, one type of sensors that could prove extremely useful in a wide variety of applications are velocity sensors. Analog velocity sensor circuits have been thoroughly investigated in the past years [1, 3, 5, 6, 9, 19, 21]. Nonetheless, researchers were unable to obtain a device that would simultaneously be compact, robust to background brightness level, insensitive to stimulus contrast and have a wide, unambiguous range of speed selectivity. Recently, novel velocity sensors that are sensitive to low contrast ....

J. Tanner and C. Mead. An integrated analog optical motion sensor. In N. York, editor, VLSI Signal Processing, II, pages 59--76. IEEE Press, 1986.

....of their component modules and their principles of operation. Data measurements obtained from the VLSI chips developed demonstrate their correct performance and their limits of operation. 1: Introduction Analog velocity sensor circuits have been thoroughly investigated in the past years [19, 2, 1, 7, 4, 17, 5]. Nonetheless, researchers were unable to obtain a device that would simultaneously be compact, robust to background brightness level, insensitive to stimulus contrast and have a wide, unambiguous, range of speed selectivity. Recently novel velocity sensors that are sensitive to low contrast ....

J. Tanner and C. Mead. An integrated analog optical motion sensor. In N. York, editor, VLSI Signal Processing, II, pages 59--76. IEEE Press, 1986.

....to sensors that incorporate the photoconversion stage and the motion processing circuitry on a single chip. The reviewed sensors are summarized in Table 1. Among the first circuits operating on a gradient algorithm was an optical mouse chip for estimating uniform image velocity in two dimensions [25] 4 . The circuit consisted of an 8 Theta8 array of velocitysensing elements. Their outputs were averaged to correct for offset effects and the aperture problem. The circuit measured velocity within a limited range for high contrasts, whereas with decreasing contrast the reported motion tended to ....

J. Tanner and C. Mead, "An Integrated Analog Optical Motion Sensor," in VLSI Signal Processing, II, S. Y. Kung, Ed. New York: IEEE Press, 1986, pp. 59-76.

....a single chip. A more detailed survey can be found in [6] We focus here on sensors of comparable size and performance to the one described in this article and assess the merits of this new scheme in relation to them. After an early implementation of a 2D velocity sensor based on a gradient method [7] showed poor performance, subsequent efforts concentrated on implementing temporal correspondence algorithms, ranging from pure correlation schemes [8] to algorithms performing correlation type motion computation on extracted image tokens [9] 10] 11] and to token based time of travel methods ....

.... schemes [8] to algorithms performing correlation type motion computation on extracted image tokens [9] 10] 11] and to token based time of travel methods [11] 12] 13] 14] Most of these previous analog VLSI motion sensors either only responded robustly to stimuli of very high contrasts [7], 11] or had an output signal that strongly depended on contrast and or illumination, as well as velocity [8] 10] 12] Another problem with some implementations [7] 9] 13] was the large size of the circuitry needed to extract a single velocity value, making them unsuitable for dense 2D ....

[Article contains additional citation context not shown here]

J. Tanner and C. Mead, "An integrated analog optical motion sensor," VLSI Signal Processing, II, S. Y. Kung, ed., pp. 59-76. New York: IEEE Press, 1986.

....that use spatiotemporal image gradients, and those that use spatiotemporal image correlation. Hardware systems that utilize each approach have been previously built. The chip reported here falls into the second category. Tanner and Mead built the first analog two dimensional optical flow chip [35][36] 30] The Tanner chip computes a single motion vector for the global two dimensional optical flow, using a gradient based scheme (for references, see [28] The motion vector is computed cooperatively by a network of circuits, each circuit using local spatiotemporal derivative information. The ....

J.E. Tanner and C. Mead, "An Integrated Analog Optical Motion Sensor," in VLSI Signal Processing, II, S.Y. Kung, Ed., New York: IEEE Press, pp. 59--76, 1986.

No context found.

J. Tanner and C.A. Mead, "An integrated analog optical motion sensor," in VLSI Signal Processing, 2, S.-Y. Kung, R. Owen, and G. Nash, Eds., p. 59 ff. IEEE Press, 1986.

No context found.

J. Tanner and C.A. Mead, "An integrated analog optical motion sensor," in VLSI Signal Processing, 2, S.-Y. Kung, R. Owen, and G. Nash, Eds., p. 59 ff., IEEE Press, 1986.

No context found.

J. Tanner and C.A. Mead. An integrated analog optical motion sensor. In S.-Y. Kung, R. Owen, and G. Nash, editors, VLSI Signal Processing, 2, page 59 ff. IEEE Press, 1986.

No context found.

J. Tanner and C. Mead. An integrated analog optical motion sensor. In R.W. Brodersen and H.S. Moscovitz, editors, VLSI Signal Processing, volume 2, pages 59-87, New York, 1988.

No context found.

J. Tanner and C. Mead, "An integrated analog optical motion sensor," in VLSI Signal Processing, II, S. Y. Kung, R. E. Owen, and J. G. Nash, Eds. New York: IEEE Press, 1986, pp. 59--76.

No context found.

J. Tanner and C. Mead, "An integrated analog optical motion sensor," in VLSI Signal Processing, II,S.-Y. Kung, R. E. Owen, and J. G. Nash, eds., pp. 59--76, IEEE, New York, 1986.

No context found.

J. Tanner and C. Mead, "An integrated analog optical motion sensor," VLSI Signal Processing, IEEE Press 2, pp. 59--76, 1986.

No context found.

J. Tanner and C. Mead, "An integrated analog optical motion sensor," VLSI Signal Processing, Vol. II, S.-Y. Kung, R. E. Owen, J. G. Nash (eds.), pp. 59-76, IEEE Press, New York, NY, 1986.

No context found.

J. Tanner and C. Mead, "An integrated analog optical motion sensor," in VLSI Signal Processing, II, S.-Y. Kung, R. E. Owen, and J. G. Nash, eds., pp. 59--76, IEEE, New York, 1986.

No context found.

J. Tanner and C. Mead, "An integrated analog optical motion sensor," in VLSI Signal Processing, II, N. York, ed., pp. 59--76, IEEE Press, 1986.

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