Trilateral constrained sparse representation for Kinect depth hole filling

Wang, Zhongyuan; Hu, Jinhui; Wang, ShiZheng; Lü, Tao

doi:10.1016/j.patrec.2015.07.025

Cited by 20 publications

(12 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After the training, independent variables in the test set are input into the model. The model learner calculates the predicted values of dependent variables and compares the predicted values with the true values to measure the actual performance of different algorithm models [27][28][29][30][31][32][33][34][35][36].…”

Section: Experiments Methodsmentioning

confidence: 99%

Modeling method of prediction model for salt field ion concentration under solar thermal system using random forest

Liu

Zhang

et al. 2019

THERM SCI

View full text Add to dashboard Cite

At present, the underground brine deposit of Lop Nor salt lake in Xinjiang, while is rich in solar energy resource and can be one high efficient solar thermal utilization area, has become an important potash production base in China. During the development of salt lake brine, the collection methods are different according to the concentration of ions in different locations. In order to solve the problems of low accuracy and high calculation cost in prediction of salt field ion concentration, a data mining method based on random forest is applied in this paper. To build the model, we collected K + , SO 4 2-, Cl-, and other two kinds of ions, among which the features included the collection time, collection locality and the number of salt pond. We used several methods to train and test the sample data, evaluated the experimental results using a variety of performance metrics and compared it with other methods at the same time. The results revealed that the optimal random forest model yielded the mean square error and coefficient of determination values of 0.073 and 0.940, which performed relatively better than support vector machine and extremely randomized trees.

show abstract

Section: Experiments Methodsmentioning

confidence: 99%

Modeling method of prediction model for salt field ion concentration under solar thermal system using random forest

Liu

Zhang

et al. 2019

THERM SCI

View full text Add to dashboard Cite

show abstract

“…a clear ambiguity exists as there is no valid depth information available for the object at all. As Table 1 illustrates, this is an incredibly rare occurrence in practice, and the hole is best left uncompleted rather than using invalid or implausible values (as per other work, [1,2,13,14,18,22,29,34,35,37,39,40,41,45,47,50,51]). Table 1 illustrates the typical occurrence frequency of the cases (1-12) on the KITTI dataset [27] (using [49] for depth estimation).…”

Section: Hole Fillingmentioning

confidence: 99%

“…Whilst many seminal color image completion techniques fall short when applied to depth maps [6,16], there are specific depth filling techniques that leverage classic inpainting approaches, with or without modifications, to fill depth values [2,23,30,51]. There have also been attempts to fill a target region in one of a set of multiview photographs [4], to fill color and depth via depth-assisted texture synthesis [46], and a myriad of approaches utilizing filters [13,14,18,34,39,41], temporal-based methods [5,25,38], reconstruction-based methods [17,36,47,50], and others [2,29,35,37,40]. We focus on the most relevant to this work [29,35,40].…”

Section: Prior Workmentioning

confidence: 99%

DepthComp: Real-time Depth Image Completion Based on Prior Semantic Scene Segmentation

Atapour-Abarghouei¹,

Breckon²

2017

Procedings of the British Machine Vision Conference 2017

View full text Add to dashboard Cite

We address plausible hole filling in depth images in a computationally lightweight methodology that leverages recent advances in semantic scene segmentation. Firstly, we perform such segmentation over a co-registered color image, commonly available from stereo depth sources, and non-parametrically fill missing depth values based on a multipass basis within each semantically labeled scene object. Within this formulation, we identify a bounded set of explicit completion cases in a grammar inspired context that can be performed effectively and efficiently to provide highly plausible localized depth continuity via a case-specific non-parametric completion approach. Results demonstrate that this approach has complexity and efficiency comparable to conventional interpolation techniques but with accuracy analogous to contemporary depth filling approaches. Furthermore, we show it to be capable of fine depth relief completion beyond that of both contemporary approaches in the field and computationally comparable interpolation strategies. 1: l ← length of the hole. 2: c ← completion case identifier. 3: if c in {1, 2, 3, 4} then 4: i ← index of leftmost pixel in the hole 5: else if c in {5, 6, 7, 8} then 6: i ← index of rightmost pixel in the hole 7: assign initial v 0 (i) according to case c 8: assign slope according to case c 9: while i is in the hole region do 10: update v(i) according to case c 11: if c in {1, 2, 3, 4} then 12: i ← i + 1.

show abstract

“…Prior work in depth hole filling [7,9,12,14,23,24] is not as comprehensive as color image completion. In the depth filling literature, there have been attempts to fill color and depth via depth-assisted texture synthesis in stereo images [15], a myriad of approaches utilizing filters [13,16], temporal-based methods [17,18], reconstruction-based methods [19,20], and others [7,9,10,21]. We focus on some of the most relevant [4,7,10,21].…”

Section: Prior Workmentioning

confidence: 99%

Extended Patch Prioritization for Depth Filling Within Constrained Exemplar-Based RGB-D Image Completion

Atapour-Abarghouei

Breckon

2018

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Citation for published item: etpourEerghoueiD emir nd frekonD oy F @PHIVA 9ixtended pth prioritiztion for depth (lling within onstrined exemplrEsed qfEh imge ompletionF9D in smge nlysis nd reognition X ISth snterntionl gonfereneD sgse PHIVD ¡ ovo de rzimD ortuglD tune PU!PWD PHIV Y proeedingsF D ppF QHTEQIRF veture notes in omputer sieneF @IHVVPAF Further information on publisher's website:Publisher's copyright statement:Additional information: Use policyThe full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that:• a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.Please consult the full DRO policy for further details.Abstract. We address the problem of hole filling in depth images, obtained from either active or stereo sensing, for the purposes of depth image completion in an exemplar-based framework. Most existing exemplar-based inpainting techniques, designed for color image completion, do not perform well on depth information with object boundaries obstructed or surrounded by missing regions. In the proposed method, using both color (RGB) and depth (D) information available from a common-place RGB-D image, we explicitly modify the patch prioritization term utilized for target patch ordering to facilitate improved propagation of complex texture and linear structures within depth completion. Furthermore, the query space in the source region is constrained to increase the efficiency of the approach compared to other exemplar-driven methods. Evaluations demonstrate the efficacy of the proposed method compared to other contemporary completion techniques.

show abstract

Trilateral constrained sparse representation for Kinect depth hole filling

Cited by 20 publications

References 29 publications

Modeling method of prediction model for salt field ion concentration under solar thermal system using random forest

Modeling method of prediction model for salt field ion concentration under solar thermal system using random forest

DepthComp: Real-time Depth Image Completion Based on Prior Semantic Scene Segmentation

Extended Patch Prioritization for Depth Filling Within Constrained Exemplar-Based RGB-D Image Completion

Contact Info

Product

Resources

About