The recognition of repetitive movements characteristic of walking people, galloping horses, or flying birds is a routine function of the human visual system. It has been demonstrated that humans can recognize such activity solely on the basis of motion information. We present a novel computational approach for detecting such activities in real image sequences on the basis of the periodic nature of their signatures. The approach suggests a low-level feature based activity recognition mechanism. This contrasts with earlier model-based approaches for recognizing such activities. 1 Introduction The motion recognition ability of the human visual system is remarkable. People are able to distinguish both highly structured motion, such as that produced by walking, running, swimming or flying birds, and more statistical patterns such as that due to blowing snow, flowing water or fluttering leaves. More subtle movement characteristics can be distinguished as well. For example, human observers ...

The recognition of nonrigid motion, particularly that arising from human movement (and by extension from the locomotory activity of animals) has typically made use of high-level parametric models representing the various body parts (legs, arms, trunk, head etc.) and their connections to each other. Such model-based recognition has been successful in some cases; however, the methods are often difficult to apply to real-world scenes, and are severely limited in their generalizability. The first problem arises from the difficulty of acquiring and tracking the requisite model parts, usually specific joints such as knees, elbows or ankles. This generally requires some prior high-level understanding and segmentation of the scene, or initialization by a human operator. The second problem, with generalization, is due to the fact that the human model is not much good for dogs or birds, and for each new type of motion, a new model must be hand-crafted. In this paper, we show that the recognition...

Many researchers interested in connectionist models accept that such models are "neurally inspired " but do not worry too much about whether their models are biologically realistic. While such a position may be perfectly justifiable, the present paper attempts to illustrate how biological information can be used to constrain connectionist models. Two particular areas are discussed. The first section deals with visual information processing in the primate and human visual system. It is argued that speed with which visual information is processed imposes major constraints on the architecture and operation of the visual system. In particular, it seems that a great deal of processing must depend on a single bottum-up pass. The second section deals with biological aspects of learning algorithms. It is argued that although there is good evidence for certain coactivation related synaptic modification schemes, other learning mechanisms, including back-propagation, are not currently supported by experimental data.

An integrated circuit that computes the velocity vector of a visual stimulus in one dimension is presented. The circuit combines optical sensors and associated electronics on a single silicon chip, processed with standard CMOS technology. The velocity is inferred from the time delay of the appearance of an image feature at two fixed locations on the chip. The circuit operates quite robustly for high-contrast stimuli over considerable irradiance and velocity ranges. With lower-contrast stimuli the output signal for a given velocity tends to decrease, while the direction selectivity is still maintained. The individual motion-sensing cells are compact, and they are therefore suited for use in dense 1D or 2D imaging arrays.

: We discuss some of the fundamental issues in the design of highly-parallel, dense, low-power motion sensors in analog VLSI. Since photoreceptor circuits are an integral part of all visual motion sensors, we discuss how the sizing of photosensitive areas can affect the performance of such systems. We review the classic gradient and correlation algorithms and give a survey of analog motion-sensing architectures inspired by them. We calculate how the measurable speed range scales with signal-tonoise ratio for a classic Reichardt sensor with a fixed time constant. We show how this speed range may be improved using a nonlinear filter with an adaptive time constant, constructed out of a diode and a capacitor, and present data from a velocity sensor based on such a filter. Finally, we describe how arrays of such velocity sensors can be employed to compute the heading direction of a moving subject and to estimate the time-to-contact between the sensor and a moving object. Keywords: motion se...

\Gamma\GammaAn integrated circuit with on-chip photoreceptors is described, that computes the bi-directional velocity of a visual stimulus moving along a given axis in the focal plane by measuring the time delay of its detection at two positions. Due to the compactness of the circuit, a dense array of such motion-sensing elements can be monolithically integrated to estimate the velocity field of an image and to extract higher-level image features through local or global interaction. Index Terms\Gamma\Gammamotion estimation, velocity sensor, optical flow, analog VLSI, robot vision. Compact Integrated Motion Sensor with Three-Pixel Interaction 3 I. Introduction A variety of image-processing tasks, such as segmentation and estimation of depth, can be considerably simplified in dynamic scenes if motion data is available. Furthermore, mobile systems rely on motion information for the computation of important parameters of ego-motion, such as time to contact and focus of expansion. Since...

This report describes research done within the Artificial Intelligence Laboratory and the Center for Biological Information Processing (Whitaker College) at the Massachusetts Institute of Technology. Support for the A.I. Laboratory 's artificial intelligence research is provided in part by the Advanced Research Projects Agency of the Department of Defense under Office of Naval Research contract N00014-85-K-0124. Support for this research is also provided by the Alfred P. Sloan Foundation, the Office of Naval Research, Cognitive and Neural Systems Division, the National Science Foundation and the McDonnell Foundation

We present a model for recovering the direction of heading of an observer who is moving relative to a scene that may contain self-moving objects. The model builds upon an algorithm proposed by Rieger and Lawton (1985), which is based on earlier work by Longuet-Higgins and Prazdny (1981). The algorithm uses velocity differences computed in regions of high depth variation to estimate the location of the .focus o.f ezpansion, which indicates the observer's heading direction. We relate the behavior of the proposed model to psychophysical observations regarding the ability of human observers to judge their heading direction, and show how the model can cope with self- moving objects in the environment. We also discuss this model in the broader context of a navigational system that performs tasks requiring rapid sensing and response through the interaction of simple task-specific routines.

The recognition of nonrigid motion, particularly that arising from human movement (and by extension from the locomotory activity of animals) has typically made use of high-level parametric models representing the various body parts (legs, arms, trunk, head etc.) and their connections to each other. Such model-based recognition has been successful in some cases; however, the methods are often difficult to apply to real-world scenes, and are severely limited in their generalizability. The first problem arises from the difficulty of acquiring and tracking the requisite model parts, usually specific joints such as knees, elbows or ankles. This generally requires some prior high-level understanding and segmentation of the scene, or initialization by a human operator. The second problem, with generalization, is due to the fact that the human model is not much good for dogs or birds, and for each new type of motion, a new model must be hand-crafted. In this paper, we show that the recognition...

How does the brain process visual information about selfmotion? In monkey cortex, the analysis of visual motion is performed by successive areas specialized in different aspects of motion processing. Whereas neurons in the middle temporal (MT) area are direction-selective for local motion, neurons in the medial superior temporal (MST) area respond to motion patterns. A neural network model attempts to link these properties to the psychophysics of human heading detection from optic flow. It proposes that populations of neurons represent specific directions of heading. We quantitatively compared single-unit recordings in area MST with single-neuron simulations in this model. Predictions were derived from simulations and subsequently tested in recorded neurons. Neuronal activities depended on the position of the singular point in the optic flow. Best responses to opposing motions occurred for opposite