Intensity-guided edge-preserving depth upsampling through weighted L 0 gradient minimization

Jung, Cheolkon; Yu, Shengtao; Kim, Joongkyu

doi:10.1016/j.jvcir.2016.11.009

Cited by 8 publications

(12 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our previous work [35], we have pointed out that depth has high edge sparsity, and to support such observations we present an example in Fig. 3.…”

Section: B Super-weighted L 0 Gradient Minimizationsupporting

confidence: 73%

“…Based on MRF model, Ferstl et al [34] designed a new regularization term by using anisotropic total generalized variation model. Jung et al [35] proposed an intensity guided weighted L 0 gradient minimization method to enhance the resolution of the depth image. Following WLS framework [36], several methods have been proposed and achieve good performance [37]- [39].…”

Section: Optimization Based Methodsmentioning

confidence: 99%

“…In our previous work [35], we have proposed intensity guided depth upsampling through weighted L 0 gradient minimization. We use L 0 gradient to construct the regularization term whose weight is computed by both intensity and depth images.…”

Section: Proposed Methodsmentioning

confidence: 99%

“…Since intensity and depth represent different information of the scene, they have inconsistent structure (e.g., adjacent pixels have similar intensities but distinct depths, or have similar depths but distinct intensities) to some extent [35]. To obtain their consistent edge structure, i.e.…”

Section: A Joint Mutual Structure Filteringmentioning

confidence: 99%

See 3 more Smart Citations

Intensity Guided Depth Upsampling Using Edge Sparsity and Super-Weighted $L_0$ Gradient Minimization

Yu¹,

Liu²,

Jung³

2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

Although depth cameras acquire depth in dynamic scenes, the captured depth images are often noisy and of low resolution. Depth images have the physical nature of being represented by smooth regions and edges in between them, i.e. depth images have high edge sparsity in the gradient domain. In this paper, we propose intensity guided depth upsampling using edge sparsity and super-weighted L 0 gradient minimization. First, we get mutual structure between intensity and depth using joint mutual structure filtering. Second, we generate an initial depth image using recursive interpolation. Next, we generate weights for L 0 gradient minimization based on gradient and entropy of the intensity image, and upsample the depth image using super-weighted L 0 gradient minimization. In super-weighted L 0 gradient minimization, we combine two main terms: Hybrid data fidelity and weighted L 0 gradient regularization. The hybrid data fidelity term combines both zero-order and first-order data differences to suppress staircase artifacts, while the weighted L 0 gradient regularization term preserves depth structures and removes noise. Finally, we further refine the depth image using adaptive fast weighted median filtering. Experiments on Middlebury and realworld scene datasets verify that the proposed method produces edge preserving depth upsampling results and outperforms state-of-the-arts in terms of both visual quality and quantitative measurements. INDEX TERMS Depth upsampling, edge sparsity, intensity guided, L 0 gradient minimization, recursive interpolation, mutual structure, weighted median filtering.

show abstract

“…In our previous work [35], we have pointed out that depth has high edge sparsity, and to support such observations we present an example in Fig. 3.…”

Section: B Super-weighted L 0 Gradient Minimizationsupporting

confidence: 73%

Section: Optimization Based Methodsmentioning

confidence: 99%

Section: Proposed Methodsmentioning

confidence: 99%

Section: A Joint Mutual Structure Filteringmentioning

confidence: 99%

See 2 more Smart Citations

Intensity Guided Depth Upsampling Using Edge Sparsity and Super-Weighted $L_0$ Gradient Minimization

Yu¹,

Liu²,

Jung³

2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

show abstract

“…Garicia et al [117] proposed a unified multi-lateral filter which can increase the image resolution in real-time. At present upsampling is an effective method to improve the resolution [118], due to the constraints in upsampling models, the high-resolution depth image obtained in this way suffers from either texture copy artifacts or depth discontinuity blur. An optimization framework proposed in [119] can tackle this problem well.…”

Section: ) Resolution Improvementmentioning

confidence: 99%

Recent Advances in 3D Data Acquisition and Processing by Time-of-Flight Camera

Chen

2019

IEEE Access

View full text Add to dashboard Cite

Three-dimensional (3D) data acquisition and real-time processing is a critical issue in an artificial vision system. The developing time-of-flight (TOF) camera as a real-time vision sensor for obtaining depth images has now received wide attention, due to its great potential in many areas, such as 3D perception, computer vision, robot navigation, human-machine interaction, augmented reality, and so on. This paper survey advances in TOF imaging technology mainly from the last decade. We focus only on recent progress of overcoming limitations such as systematic errors, object boundary ambiguity, multipath error, phase wrapping, and motion blur, and address the theoretical principles and future research trends as well. INDEX TERMS Vision sensor, time of flight, depth image, sensing device, computer vision, 3D vision.

show abstract

Texture-guided depth upsampling using Bregman split: a clustering graph-based approach

2018

View full text Add to dashboard Cite

Intensity-guided edge-preserving depth upsampling through weighted L 0 gradient minimization

Cited by 8 publications

References 56 publications

Intensity Guided Depth Upsampling Using Edge Sparsity and Super-Weighted $L_0$ Gradient Minimization

Intensity Guided Depth Upsampling Using Edge Sparsity and Super-Weighted $L_0$ Gradient Minimization

Recent Advances in 3D Data Acquisition and Processing by Time-of-Flight Camera

Texture-guided depth upsampling using Bregman split: a clustering graph-based approach

Contact Info

Product

Resources

About