weighted joint bilateral filter with slope depth compensation filter for depth map refinement

38
Weighted Joint Bilateral Filter with Slope Depth Compensation Filter for Depth Map Refinement Takuya Matsuo, Norishige Fukushima and Yutaka Ishibashi VISAPP 2013 International Conference on Computer Vision Theory and Application 1

Upload: kanan

Post on 09-Feb-2016

82 views

Category:

Documents


7 download

DESCRIPTION

Weighted Joint Bilateral Filter with Slope Depth Compensation Filter for Depth Map Refinement. Takuya Matsuo, Norishige Fukushima and Yutaka Ishibashi VISAPP 2013 International Conference on Computer Vision Theory and Application. Outline. Introduction Related Works Proposed Method - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

1

Weighted Joint Bilateral Filter with Slope Depth Compensation Filter

for Depth Map Refinement

Takuya Matsuo, Norishige Fukushima and Yutaka Ishibashi

VISAPP 2013 International Conference on Computer Vision Theory and

Application

Page 2: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

2

Outline

• Introduction• Related Works• Proposed Method–Weighted Joint Bilateral Filter – Slope Depth Compensation Filter

• Experimental Results• Conclusion

Page 3: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

3

INTRODUCTION

Page 4: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

4

Introduction

• Goal : Using two filters to get more accurate disparity map in real-time.

• Consideration– Noise reduction – Correct edges – Blurring control

Goal

Page 5: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

5

RELATED WORKS

Page 6: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

6

Left Image Right Image

Related Works

• Stereo Matching

Page 7: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

Related Works

Local Global

Estimate accuracy Low High

Calculation cost Low High

Methods Pixel matching Block matching

(Optimization methods) Graph cuts

Belief propagation

Example

Page 8: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

Related Works

• Flow Chart (Local)

1• Matching Cost Computation

2• Cost Aggregation

3• Disparity Map Computation/Optimization

4• Disparity Map Refinement• Disparity Map Refinement

Page 9: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

9

Related Works

• Depth map refinement with filter– Median filter– Bilateral filter

Input depth map Output depth map Filter

Page 10: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

10

Related Works

• Bilateral filter – Space weight:Near pixels has large weight – Color weight:Similar color pixels has large weight

• Smoothing – Keep edges –Weak in spike noise

Page 11: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

11

Related Works

• Joint bilateral filter – Add in the reference image – Color weight is calculated by the reference – Keep object edges of the reference

Reference : Low noise Target : High noise Filtered image

Page 12: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

12

Related Works

• Joint bilateral filter– Noise reduction O– Correct edge O– Blurring X • Mixed depth values • Spreading error regions

• Multilateral filter– Space + Color + Depth– Boundary recovering X

Page 13: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

13

PROPOSED METHOD

Page 14: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

14

Proposed Method

• Weighted joint bilateral filter – Noise reduction – Edge correction

• Slope depth compensation filter – Blurring control

Page 15: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

15

Weighted Joint Bilateral Filter

– 𝐷: Depth value – 𝑝: Coordinate of current pixel – 𝑠: Coordinate of support pixel – 𝑁: Aggregation set of support pixel – 𝑤(), (): Space/color weight 𝑐– 𝜎𝑠,𝜎𝑐: Space/color Gaussian distribution – 𝑅𝑠: Weight map

Page 16: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

16

Weighted Joint Bilateral Filter

• Add in the weight map – Controlling amount of influence on a pixel –Weight of the edge and error is small

Joint bilateral filter 𝜎- Mixed depth values 𝜎- Spreading error regions

Page 17: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

17

Weighted Joint Bilateral Filter

• Making weight map – Space/color/disparity weight – Sum of nearness of space,

color, and disparity between center pixel and surrounding pixels.

Page 18: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

18

Weighted Joint Bilateral Filter

• Mask image is made by Speckle Filter– Detecting speckle noise –Weight of speckle region is 0

Red region: speckle noise Weight = 0

Page 19: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

19

Weighted Joint Bilateral Filter

Page 20: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

20

Slope Depth Compensation Filter

• Weighted joint bilateral filter– Remaining small blurring – Difference between foreground and background

color is small

• Slope depth compensation filter – Reason of blurring is mixed depth value – Convert mixed value to non-blurred candidate

using initial depth map

Removing remaining blur

Page 21: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

21

Slope Depth Compensation Filter

– X in Dx {INITIAL;WJBF;SDCF}∈

Page 22: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

22

Slope Depth Compensation Filter

Page 23: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

23

Proposed Method

Initial disparityStereo matching

Noise reduction/ edge correction Weighted Joint Bilateral F.

Blurring control Slope Depth Compensention F.

Page 24: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

24

EXPERIMENTAL RESULTS

Page 25: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

25

Experimental Results

• Evaluating accuracy improvement for various types of depth maps – Block Matching (BM) – Semi-Global Matching (SGM) – Efficient Large-Scale (ELAS) – Dynamic Programing (DP) – Double Belief Propagation (DBP)

Page 26: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

26

Experimental Results

Page 27: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

27

Experimental Results

Page 28: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

28

Page 29: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

29

Page 30: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

30

Page 31: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

31

Experimental Results

• Comparing proposed method with cost volume refinement(Teddy).

Yang, Q., Wang, L., and Ahuja, N. A constantspace belief propagation algorithm for stereo matching.In Computer Vision and Pattern Recognition(2010).

32 times slower

Page 32: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

32

Experimental Results

Page 33: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

33

Experimental Results

• Device : Intel Core i7-920 2.93GHz• Comparing running time (ms) of BM plus proposed

filter with selected stereo methods.

Page 34: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

34

Experimental Results

Page 35: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

35

Experimental Results

• Use the proposed filter for depth maps from Microsoft Kinect.

Page 36: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

36

Page 37: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

37

CONCLUSION

Page 38: Weighted Joint Bilateral Filter with Slope Depth Compensation Filter  for Depth Map Refinement

38

Conclusion

Contribution• The proposed methods can reduce depth noise and

correct object boundary edge without blurring. • Amount of improvement is large when an input

depth map is not accurate.

Future Works• Investigating dependencies of input natural images

and depth maps.