Optimal Mode Selection For Robust Video Transmission - PowerPoint PPT Presentation

1 / 11
About This Presentation
Title:

Optimal Mode Selection For Robust Video Transmission

Description:

Optimal Mode Selection For Robust Video Transmission. EE368C Project Proposal ... High Complexity Mode Selection : Intra mode if (A (SAD(x,y) 500) ... – PowerPoint PPT presentation

Number of Views:51
Avg rating:3.0/5.0
Slides: 12
Provided by: sangoh
Category:

less

Transcript and Presenter's Notes

Title: Optimal Mode Selection For Robust Video Transmission


1
Optimal Mode Selection For Robust Video
Transmission
  • EE368C Project Proposal
  • Stanford University
  • Sangoh Jeong

Feb.1, 2001
2
Outline
  • Goals
  • Existing Mode Decision Processes
  • R-D Based Mode Selection
  • Proposed New Coding Mode
  • Proposed Experiment
  • Project Schdule

3
Goals
  • Optimal Switching of Coding Modes
  • Proposes an additional coding mode
  • Considers the trade-off between coding efficiency
    and error resiliency
  • Improves PSNR characteristic in error-prone
    environment
  • Uses an intelligent encoder without feedback
  • Decides the optimal mode with moderate complexity

4
Mode Decision Processes
  • Low Complexity Mode Selection
  • Error-Free Environment
  • Error-Prone Environment
  • High Complexity Mode Selection

Intra mode if (A lt (SAD(x,y) 500))
  • Minimize Lagrangian Cost Function

5
Lagrangian Cost Function
  • D SSD between the original MB and its Rec.
  • R bitrate associated with choosing Mode

Rate R
Lines of constant J
Distortion D
6
Proposed Mode Selection
  • Cost Function of Each Mode
  • Intra Mode
  • Inter Mode
  • In-Between Mode
  • Value of
  • Choose the Mode which has the mimimum .

7
Operation Of In-Between Mode
  • Prediction Error
  • Leaky Prediction
  • Attenuates the energy of the prediction signal
  • Superimposed errors decay over time
  • Intra perfect, Inter weak

MCed Frame
Intra Frame
MB1
MB2
8
Proposed Experiment
9
Experiment Conditions
  • Types of Trasmission Errors to be used
  • Burst Errors
  • Random loss of 1 GOB in every two frame
  • Random Bit Errors
  • Randomize 5 of bits in each frame
  • Encoder Software to be used
  • H.263 (tmn10) encoder
  • Mode decision
  • - Low Complexity Uses threshold
  • - High Complexity Uses Lagrangian cost
    function

10
Expected Result
0
Two-mode switching
time(s)
Loss in signal-to-noise ratio of decoded video
signal after errors happen
11
Project Schdule
Write a Comment
User Comments (0)
About PowerShow.com