1 Scope
1.1 Application
1.2 Limitations
1 Introduction
2 BTFR
3 Detectors
3.1 Input conversion
3.2 Crop and offset
3.3 Matching
3.3.1 Matching
statistics
3.3.2 MPSNR
3.3.3 Matching vectors
3.4 Spatial frequency analysis
3.4.1 Pyramid transform
3.4.2 Pyramid SNR
3.5 Texture analysis
3.6 Edge analysis
3.6.1 Edge detection
3.6.2 Edge differencing
3.7 MPSNR analysis
4 Integration
5 Registration
6 References
1 Introduction
2 Objective measurement of video quality
based on edge degradation
2.1
Edge
PSNR (EPSNR)
2.2 Post adjustments
2.2.1 De-emphasis
of high EPSNR
2.2.2 Considering
blurred edges
2.2.3 Scaling
2.3 Registration accuracy
2.4 The block diagram of the model
3 Objective data
4 Conclusion
5 Reference
1 Introduction
2 General description of the IES system
3 Correction of offset and gain
3.1 Temporal offset
3.2 Spatial offset
3.3 Gain
4 Image segmentation
4.1 Plane regions
4.2 Edge regions
4.3 Texture regions
5 Objective measurement
6 Database of impairment models
7 Estimation of impairment models
7.1 Computation of Wi
7.2 Computation of Fi and Gi
8 References
1 Introduction
2 Normative reference
3 Definitions
4 Overview of the VQM computation
5 Sampling
5.1 Temporal indexing of original and
processed video files
5.2 Spatial indexing of original and
processed video frames
5.3 Specifying rectangular sub-regions
5.4 Considerations for video sequences
longer than 10 s
6 Calibration
6.1 Spatial registration
6.1.1 Overview
6.1.2 Interlace issues
6.1.3 Required inputs
to the spatial registration algorithm
6.1.4 Sub-algorithms
used by the spatial registration algorithm
6.1.5 Spatial registration using arbitrary scenes
6.1.6 Spatial
registration of progressive video
6.2 Valid region
6.2.1 Core valid region
algorithm
6.2.2 Applying the core
valid region algorithm to a video sequence
6.2.3 Comments on valid
region algorithm
6.3 Gain and offset
6.3.1 Core gain and
level offset algorithm
6.3.2 Using scenes
6.3.3 Applying gain and level offset corrections
6.4 Temporal registration
6.4.1 Frame-based
algorithm for estimating variable temporal delays between original and
processed video sequences
Step 1: Calibrate the video sequences
Step 2: Select
the sub-region of video to be used
Step 3: Spatially
sub-sample the original and processed images
Step 4: Normalize
the sub-sampled images
Step 5: Compare
processed images to original images
Step 6: Perform
an overall check for still video
Step 7: Temporally
register each processed image
Step 8: Perform
a stillness check on each processed image
Step 9: Form
a histogram of all defined temporal registrations
Step 10: Form a smoothed histogram
Step 11: Examine
the histogram information
6.4.2 Applying temporal
registration correction
7 Quality features
7.1 Introduction
7.1.1 S-T regions
7.2 Features based on spatial gradients
7.2.1 Edge enhancement
filters
7.2.2 Description of
features fSI13
and fHV13
7.3 Features based on chrominance
information
7.4 Features based on contrast information
7.5 Features based on ATI
7.6 Features based on the cross product of
contrast and ATI
8 Quality parameters
8.1 Introduction
8.2 Comparison functions
8.2.1 Error ratio and
logarithmic ratio
8.2.2 Euclidean
distance
8.3 Spatial collapsing functions
8.4 Temporal collapsing functions
8.5 Non-linear scaling and clipping
8.6 Parameter naming convention
8.6.1 Example parameter
names
9 General model
10 References
2 Video
materials
2.1 SRC and HRC
1 Methodology
for the evaluation of objective model performance
4 Evaluation of results
5 PSNR data
6 References