Motion detection A computational problems: Local motion signals are ambiguous How to combine motion signals into one

e coherent whole Computation of Visual Motion Aperture: An opening that allows only a partial view of an object Correspondence problem (motion): The problem faced by the motion detection system of knowing which feature in frame 2 corresponds to which feature in frame 1 Aperture problem: The fact that when a moving object is viewed through an aperture (or a receptive field), the direction of motion of a local feature or part of an object may be ambiguous The correspondence problem

The aperture problem

Motion information from several local apertures (or receptive fields) can be combined to determine the global motion of the object There are several directions of motion within each aperture that are compatible with the stimulation the receptor is receiving Whichever possible motion direction is the same in all apertures is the true global motion direction of the object Building a global-motion detector

Computation of Visual Motion We can say something about where global-motion detectors are: Lesions in magnocellular layers of LGN impair perception of large, rapidly moving objects Middle temporal lobe (MT) also plays an important role in motion perception The vast majority of neurons in MT are selective for motion in a particular direction Motion-sensitive areas in the brain (Part 1)

Computation of Visual Motion Newsome and Pare (1988) conducted a study on motion perception in monkeys Trained monkeys to respond to correlated dot motion displays The MT area of the monkeys was lesioned

Result: Monkeys needed about ten times as many dots to correctly identify direction of motion The Newsome and Pare paradigm

Computation of Visual Motion Interocular transfer: The transfer of an effect (such as adaptation) from one eye to another MAE exhibits interocular transfer Therefore, MAE must occur in neurons that respond to both eyes Input from both eyes is combined in area V1, so MAE must be in V1 or later Recent studies with fMRI confirm that adaptation in MT is responsible for MAE Motion Information for Form A. Motion alone can indicate the 3-D form of objects. B. Originally, this was called the kinetic depth effect. C. More generally, these abilities are called structure-from-motion (SFM) -- referring to the extraction of object structure from information in moving displays Structure From Motion Moving dots are often used to study SFM. When stationary dots do not reveal the contours or surface information of objects. Thus, they can be used to study the motion effects independent of surface properties. Mathematical Analyses of SFM Focus has been on Rigid Motion. Rigid motion is the geometric term describing motion of an object in space during which there are no changes in the distances between any two points on the object. In other words, the object does not bend or deform during motion. Mathematical Analyses of SFM

Focus on Rigid Motion

Mathematical Analyses of SFM: Rigid Motion The Problem of Determining SFM The problem of SFM in rigid motion is for the observer to recover 3-D form from the changing 2-D projection of an object. A number of theorems have been proven showing that from a small number of points (3-5) moving in the 2-D image, the 3-D structure of a rigid object can be recovered The Problem of Determining SFM Some have argued that the visual system will fit a rigid solution to motion displays whenever this is possible. Non-Rigid Motion We are clearly able to see non-rigid shape. Examples: Your hand (jointed motion) Jellyfish (elastic motion) Non-rigid motion Much research has looked at point-light walker displays, in which perceivers see a person walking from motion of only a few points of light. (Johansson) This is also called biological motion. Using Motion Information Biological motion: The pattern of movement of all Animals Biological Motion Humans are very sensitive easily identify various actions actor traits (gender, identity) Importantly: - affective (emotional) information - communicative information

Mathematical analysis of non-rigid motion is much harder What constraints are there? What distinguishes unified, non-rigid motion from disconnected motion of dots? These are unsolved problems! Using Motion Information Avoiding imminent collision: How do we estimate the time to contact (TTC) of an approaching object? Tau (): Information in the optic flow that could signal TTC without the necessity of estimating either absolute distances or rates The ratio of the retinal image size at any moment to the rate at which the image is expanding is tau, and TTC is proportional to tau: X/X With enough training, it seems even possible to do it completely without any influence from consciousness: samurai halving a bullet But less experienced observers (or less observant observers) do not react fast enough. Using Motion Information How do we use motion information to navigate? Optic array: The collection of light rays that interact with objects in the world in front of a viewer Optic flow: The changing angular position of points in a perspective image that we experience as we move through the world The optic flow field produced by movement forward in space

Using Motion Information Focus of expansion (FOE): The point in the center of the horizon from which, when we are in motion, all points in the perspective image seem to emanate This is one aspect of optic flow The focus of expansion tells the observer which way they are heading Only so when no gaze shifts are made

Information for Motion Perception There are multiple sources of information or multiple situations that lead to perceived motion. 1. RETINAL DISPLACEMENT is the changing position of an object's image on your eye. 2. OPTICAL PURSUIT occurs when we track a moving object with our eye. The image stays on the fovea, yet we perceive it as moving. Eye Movements Why do we perceive the pencil to be in motion in the first case, but perceive the dot to be stationary in the second case? After all, both items moved across our retinas to the left Because in the second case there was an eye Movement Eye Movements Types of eye movements: Smooth pursuit: The eyes smoothly follow a moving target Saccade: A rapid movement of the eyes that changes fixation from one object or location to another Vergence: The two eyes move in opposite directions, as when both eyes turn towards the nose Reflexive: Automatic and involuntary eye Movements

Eye Movements Saccadic suppression: The reduction of visual sensitivity that occurs when we make saccadic eye movements Saccadic suppression eliminates the smear from retinal image motion during an eye movement Self-motion correction Efference copy Eye Movements How do we discriminate motion across the retina that is due to eye movements vs. object movements? Comparator: An area of the visual system that receives one copy of the order issued by the motor system when the eyes move (the other copy goes to the eye muscles) The comparator can compensate for the image changes caused by the eye movement