Open Problems in Computer Vision(oProCV) Topic 4. Motion MSc. Luis A. Mateos.

Open Problems in Computer Vision(oProCV)

Topic 4. Motion

MSc. Luis A. Mateos


Motion

detecting motionoptical flowtracking single objectsshadows and occlusionsdynamic scene representations

Motion means a change in the location of a body. A body which does not move is said to have constant (time-invariant) position.

Motion is always observed and measured relative to a frame of reference.

Everything in the universe is moving[1].

Motion signifies any spatial and/or temporal change in a physical system.

[1] De Grasse Tyson, N., Liu, C., & Irion, R. 2000. One Universe: At home in the cosmos. p.20–21. Joseph Henry Press. ISBN 0-309-06488-0


Motion

http://www.nap.edu/html/oneuniverse/motion_20-21.html














Motion detection quantifies the motion of a moving object within the field of view.

Algorithm by a fixed camera compares the current image with a reference image and simply counts the number of different pixels.

Images will differ due to factors such as varying lighting, camera flicker, etc..(noise),so pre-processing is needed to reduce the number of false positive.


Motion Detection

Optical Flow

Is the pattern of apparent motion of objects, surfaces, and edges in a visual scene caused by the relative motion between an observer and the scene.

Try to calculate the motion between two image frames which are taken at times t and t + δt at every pixel position.

For a 2D+t dimensional case (3D or n-D cases are similar) a pixel at location (x,y,t) with intensity I(x,y,t) will have moved by δx, δy and δt between the two image frames,


Optical Flow

Every arrow from the optical flow is a vector, defined by a magnitude and an angle.


Optical Flow


Tracking single objects

Considerations

Type of object:- Rigid- Deformable

How to isolate the object from:- The background- Other objects- Noise (Ilumination, camera,)


Tracking single objects

Common algorithms

Blob tracking: Segmentation of object interior (for example blob detection, block-based correlation or optical flow).

Kernel-based tracking (Mean-shift tracking): An iterative localization procedure based on the maximization of a similarity measure.

Contour tracking: Detection of object boundary (e.g. active contours or Condensation algorithm)

Visual feature matching: Registration


Shadows and occlusions

ShadowsA shadow is an area where direct light from a light source cannot reach due to obstruction by an object.

It occupies all of the space behind an opaque object with light in front of it.

The cross section of a shadow is a two-dimensional silhouette, or reverse projection of the object blocking the light.



Occlusions Partial occlusion, when the object hide only part of it from the camera.



Occlusions Complete occlussion, when the object is hidden, in the current frame it cannot be found.


Dynamic scene representations

One of the more powerful impressions created by vision is that of a coherent, richly detailed world where everything is present simultaneously.

Indeed, this impression is so compelling that we tend to ascribe these properties not only to the external world, but to our internal representations as well.

For example, changes in images of real-world scenes often go unnoticed when made during a saccade, flicker, blink, or movie cut.

This "change blindness" provides strong evidence against the idea that our brains contain a picture-like representation of the scene that is everywhere detailed and coherent.


Dynamic scene representations

How then do we represent a scene?

It is argued here that focused attention provides spatiotemporal coherence for the stable representation of one object at a time.

It is then argued that the allocation of attention can be co-ordinated to create a "virtual representation".

In such a scheme, a stable object representation is formed whenever needed, making it appear to higher levels as if all objects in the scene are represented in detail simultaneously[2].

[2] The Dynamic Representation of ScenesAuthor: Rensink, Ronald A.Source: Visual Cognition, Volume 7, Numbers 1-3, 1 January 2000 , pp. 17-42(26)Publisher: Psychology Press, part of the Taylor & Francis Group


Thank you

Open Problems in Computer Vision(oProCV) Topic 4. Motion MSc. Luis A. Mateos.

Documents

Transcript of Open Problems in Computer Vision(oProCV) Topic 4. Motion MSc. Luis A. Mateos.