19
Noisy Video Super- Resolution Feng Liu, JinjunWang,ShenghuoZhu (MM’08) University of Wisconsin-Madison, NEC Laboratories America, Inc. 第第 第第第 一: B95902105 第第第 第第第R98922046 第第第第 R97944012

Noisy Video Super-Resolution

  • Upload
    terra

  • View
    31

  • Download
    0

Embed Size (px)

DESCRIPTION

Feng Liu, JinjunWang,ShenghuoZhu (MM’08) University of Wisconsin-Madison, NEC Laboratories America, Inc. Noisy Video Super-Resolution. 第一組: 資訊四 B95902105 黃彥達 資訊碩一 R98922046 蔡旻光 網媒碩二 R97944012 鄒志鴻. Outline. Introduction Goal File Format Noise Reduced Image - PowerPoint PPT Presentation

Citation preview

Page 1: Noisy Video Super-Resolution

Noisy Video Super-Resolution

Feng Liu, JinjunWang,ShenghuoZhu (MM’08)University of Wisconsin-Madison, NEC Laboratories America, Inc.

第一組: 資訊四 B95902105 黃彥達 資訊碩一 R98922046 蔡旻光 網媒碩二 R97944012 鄒志鴻

Page 2: Noisy Video Super-Resolution

2

Outline

Introduction Goal File Format

Noise Reduced Image Proposed Approach Motion Estimation & Estimated Super-

Resolution Result Implementation Result Conclusion

Page 3: Noisy Video Super-Resolution

3

Introduction

Low-quality videos often not only have limited resolution but also suffer from noise In fact, the requirements of de-noising & super-

resolution is quite similar

This paper present a unified framework which achieves simultaneous video de-noising and super-resolution algorithm by some measurements of visual quality

Page 4: Noisy Video Super-Resolution

Goal

Refine low-quality videos from YouTube, and make the video better effects, which has better quality by human eyes.

Input is low-quality and noise-included (block effects or somewhat noise) videos

Page 5: Noisy Video Super-Resolution

Noise-Reduced Image

mv-SAD Gaussian-space

Gaussian-time

| p(I,j) – p(i’, j’) | > threshold

Page 6: Noisy Video Super-Resolution

Gaussian Space

Frame t

Pixel(I,j)

Standard deviation

Set Mean = 0

Page 7: Noisy Video Super-Resolution

Motion Vector

Frame tPixel ( i , j , t)

Frame t+1

Pixel ( i + mv_i , j + mv_j , t+1)

(mv_i , mv_j)

Page 8: Noisy Video Super-Resolution

Gaussian Time

Fram

e t

- 2Fram

e t

- 1Fram

e t

Space Gaussian

Time Gaussian

Pixel(I,j)

Fram

e

t+1Fram

e

t+2Fram

e t

Page 9: Noisy Video Super-Resolution

Noise-Reduced ImageBefore After

Page 10: Noisy Video Super-Resolution

10

Proposed Approach – 1 / 4 Consider the visual quality with respect to

the following 3 aspects: Fidelity Preserving▪ To achieve similar high-resolution result

Detail Preserving▪ Enhanced details (edge)

Spatial-Temporal Smoothness▪ Remove undesirable high-frequency contents (e.g. jitter)

Page 11: Noisy Video Super-Resolution

11

Proposed Approach – 2 / 4 Fidelity Preserving

Conventional metrics:▪ Measure fidelity by the difference between Ih & Il would

be problematic & waste useful time-space information in video

Proposed metrics:▪ Estimate an approximation of super-resolution results

from space-time neighboring pixels▪ The fidelity measurement:

see next page for details

noised

Page 12: Noisy Video Super-Resolution

12

Proposed Approach – 3 / 4 Detail Preserving

Enhanced details (edge)

Contrast preserving▪ Human visual system is more sensitive to contrast

than pixel values▪ Gradient fields of Ih & should be close

,where Wk is one or zero if the patchk with/o edges (canny detector)

Page 13: Noisy Video Super-Resolution

13

Proposed Approach – 4 / 4 (Spatial-Temporal) Smoothness

Smooth results are often favored by the human system Encourage to minimize:

A 2-D Laplace filter may be

Spatial-temporal Laplacian

OR

Page 14: Noisy Video Super-Resolution

14

An Optimization Problem

Proposed Measurements

A quadratic minimization problem to solve (AX = b):

Contrast

Similarity

Detail Information(edge)

Spatial-Temporal Smoothness

Page 15: Noisy Video Super-Resolution

15

Implementation – 1 / 2

InputlowhI~

hI~

hG~

hI~

)1,()1,(~~

tIwtIw hmv

hmv

= X

I

6 -1 … -1-1 6 -1 … -1 -1 6 -1 … -1

Laplacian

Gradient-1 0 1 … 1 -1 0 1 … 1 -1 0 1 … 1

EdgeMinimize

Motion Estimation

+

Result (X)

Fidelity

Bilateral filter

mvw

fidv

fidg

dt

sm

Page 16: Noisy Video Super-Resolution

16

Implementation – 2 / 2

Adjustments for the weight terms The measurement term is more emphasized if

the weight is larger

By iteratively experiments for our test data, we took

However, we found that for different videos, the best weight sets may be also different

3.0,1.0,1,1 smdtfidgfidv

Page 17: Noisy Video Super-Resolution

17

Result

352 x 288 Result

Page 18: Noisy Video Super-Resolution

18

Conclusion

The proposed framework formulates noisy video super-resolution as an optimization problem, aiming to maximize the visual quality of the result

The measurements of fidelity-preserving, detail-preserving and smoothness are considered to maximize the visual quality results

Page 19: Noisy Video Super-Resolution

19

Thank you!!