motion representation

11
Motion representation Global: Entire motion field is represented by a few global parameters Block-based: Entire frame is divided into blocks, and motion in each block is characterized by a few parameters. Pixel-based: One MV at each pixel, with some smoothness constraint between adjacent MVs. Region-based: Entire frame is divided into regions, each region corresponding to an object or subobject with consistent motion, represented by

Upload: ulmer

Post on 22-Mar-2016

36 views

Category:

Documents


0 download

DESCRIPTION

Motion representation. Pixel-based: One MV at each pixel, with some smoothness constraint between adjacent MVs. Global: Entire motion field is represented by a few global parameters. Region-based: Entire frame is divided into regions, each region corresponding - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Motion representation

Motion representation

Global:Entire motion field isrepresented by a fewglobal parameters

Block-based:Entire frame is dividedinto blocks, and motionin each block ischaracterized by a fewparameters.

Pixel-based:One MV at each pixel,with some smoothnessconstraint betweenadjacent MVs.

Region-based:Entire frame is dividedinto regions, eachregion correspondingto an object or subobjectwith consistentmotion, represented bya few parameters.

Page 2: Motion representation

Typical camera movements

Page 3: Motion representation

Backward and forward motion estimation

Page 4: Motion representation

Block matching

Page 5: Motion representation

Typically, the displaced frame difference, computed on a block-by-block basis, is minimized. How?

- exhaustive search, or (much faster)

- gradient-based, multiresolution, or log search

Real MV may not always be multiples of pixels. To allow subpixel MV, the search step size must be less than 1 pixel, e.g. ½ pixel (the image has to be interpolated…)

Page 6: Motion representation

Mesh-based motion estimation

Using a triangular or a quadrilateral mesh

Page 7: Motion representation

Block-based vs mesh-based

In mesh-based, motion in each element is interpolated from nodal MVs

In mesh-based: two problems

– Set up the mesh in the anchor frame, so that the mesh conforms with object boundaries: mesh generation

– Given a mesh in the anchor frame, determine nodal positions in the target frame: motion estimation

Page 8: Motion representation

Block-based vs mesh-based

Page 9: Motion representation

Video coding: outline

• Predict a new frame from a previous frame and only code the prediction error

• Prediction error will be coded using the DCT method• Those regions that cannot be predicted well will be coded

directly using DCT• Work on each macroblock (MB) (16x16 pixels) independently

for reduced complexity – Motion compensation done at the MB level – DCT coding of error at the block level (8x8 pixels)

Page 10: Motion representation

GOP

Page 11: Motion representation

Coder and decoder