Specular Reflections Removal for Endoscopic Image Sequences With Adaptive-RPCA Decomposition

Li, Ranyang; Pan, Junjun; Si, Yaqing; Yan, Bin; Hu, Yong; Qin, Hong

doi:10.1109/tmi.2019.2926501

Cited by 48 publications

(44 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…from Theorem 1 and Corollary 1 of [24] we get directly (6). Notice that (6) generalizes the expressions for the variational form of the nuclear norm [25], which are obtained by setting p = 1 in (6) (recall that the nulcear norm is the Schatten-1 norm).…”

Section: Problem Formulation and Batch Irls Rpcamentioning

confidence: 75%

“…S INCE the seminal work of Candes et al [1], robust principal component analysis (RPCA) has attracted considerable attention both in research and in applications. Following the principal component pursuit (PCP) [1], a plethora of new algorithms have been developed [2] and applied in diverse fields such as low-level image processing [3], [4], medical imaging [5], [6], hyperspectral imaging [7], [8], video processing and analysis [9], [10], computer vision [11], [12] to name but a few [13], [14].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Online Reweighted Least Squares Robust PCA

Rontogiannis

Giampouras

Koutroumbas

2020

IEEE Signal Process. Lett.

View full text Add to dashboard Cite

The letter deals with the problem known as robust principal component analysis (RPCA), that is, the decomposition of a data matrix as the sum of a low-rank matrix component and a sparse matrix component. After expressing the low-rank matrix component in factorized form, we develop a novel online RPCA algorithm that is based entirely on reweighted least squares recursions and is appropriate for sequential data processing. The proposed algorithm is fast, memory optimal and, as corroborated by indicative empirical results on simulated data and a video processing application, competitive to the state-of-the-art in terms of estimation performance.

show abstract

Section: Problem Formulation and Batch Irls Rpcamentioning

confidence: 75%

Section: Introductionmentioning

confidence: 99%

Online Reweighted Least Squares Robust PCA

Rontogiannis

Giampouras

Koutroumbas

2020

IEEE Signal Process. Lett.

View full text Add to dashboard Cite

show abstract

“…Firstly, from 16 pieces of literature [ 17 , 18 , 21 – 23 , 25 , 26 , 28 , 31 , 32 , 34 – 39 ] concerning SR elimination from endoscopic images, we extracted 50 images accompanied by their corresponding result images as a new dataset, called Ref_set. Due to the limited research specializing in SR regions elimination of colposcopic images, we extended the images to a broader range of endoscopic images, including some colonoscopic and laparoscopic images.…”

Section: Resultsmentioning

confidence: 99%

“…It is worth noting that colposcopic images have many similarities with other endoscopic images, so some standard research methods can be applied to the problem of eliminating SR regions in the colposcopic images. Colonoscopy [ 25 ], thoracoscopy [ 26 ], laparoscopy [ 27 ], and more researches on eliminating SR regions in these images [ 28 , 29 ] can be used for reference. In the report of Wang et al [ 30 ], they used Arnold's method [ 25 ] in colonoscopy images and combined it with the exemplar-based method to eliminate SR regions in colposcopic images, showing better effect.…”

Section: Related Workmentioning

confidence: 99%

Integration of Global and Local Features for Specular Reflection Inpainting in Colposcopic Images

Wang

et al. 2021

Journal of Healthcare Engineering

View full text Add to dashboard Cite

Objective. To explore an inpainting method that can balance texture details and visual observability to eliminate the specular reflection (SR) regions in the colposcopic image, thus improving the accuracy of clinical diagnosis for cervical cancer. Methods. (1) To ensure smoothness, Gaussian Blur and filling methods are applied to the global image. (2) Striving to preserve the anatomical texture details of the colposcopic image as much as possible, the exemplar-based method is applied to local blocks. (3) The colposcopic images inpainted in the previous two steps are integrated, so that important information of non-SR regions is preserved based on eliminating SR regions. Results. In the subjective visual assessment of inpainting results, the average of 3.55 ranks first in the five comparison sets. As to the clinical test, comparing the diagnosis results of 6 physicians before and after eliminating SR regions, the average accuracy of two kinds of classifications increased by 1.44% and 2.03%, respectively. Conclusions. This method can effectively eliminate the SR regions in the colposcopy image and present a satisfactory visual effect. Significance. As a preprocessing method for computer-aided diagnosis systems, it can also improve physicians’ accuracy in clinical diagnosis.

show abstract

“…Recently researches have shown that the PCP method has strong theoretical guarantees for the video foregroundbackground separation [18], endoscopic images highlight removing and so on. For example, Li et al [22] proposed Adaptive-RPCA which can remove the specular reflections in endoscopic image sequences well. however, researchers also found that the PCP method has many defects [19], [20] as follows: 1) the nuclear norm in the PCP method is the l 1 -norm of the singular value essentially, which processes the singular value uniformly, and always generates the problem of over punishment of the singular value; 2) the calculation cost of the PCP method is relatively high; 3) the surrogate function in the PCP method is a biased estimate which has low approximation accuracy.…”

Section: Introductionmentioning

confidence: 99%

Foreground-Background Separation via Generalized Nuclear Norm and Structured Sparse Norm Based Low-Rank and Sparse Decomposition

Yang

et al. 2020

IEEE Access

View full text Add to dashboard Cite

Low-rank and sparse decomposition (LRSD) has attracted wide attention in video foregroundbackground separation and many other fields. However, the traditional LRSD methods have many tough problems, such as the problems of the low accuracy of the surrogate functions of rank and sparsity, ignoring the spatial information of the videos and sensitivity to noise, etc. To deal with these problems, this paper proposes the generalized nuclear norm and structured sparse norm (GNNSSN) method based LRSD for video foreground-background separation, which introduces the generalized nuclear norm (GNN) and the structured sparse norm (SSN) to approximate the rank function and the l 0 -norm of the LRSD method. In addition, we extend our proposed model to a robust model against noise for practical applications, and we called the extended method as the robust generalized nuclear norm and structured sparse norm (RGNNSSN) method. At last, we use the alternating direction method of multipliers (ADMM) to solve our proposed two methods. Experimental results and discussions on video foreground-background separation demonstrate that our proposed two methods have better performances than other LRSD based foreground-background separation methods.INDEX TERMS Low-rank and sparse decomposition, generalized nuclear norm, structured sparse norm, alternating direction method of multipliers, foreground-background separation.

show abstract

Specular Reflections Removal for Endoscopic Image Sequences With Adaptive-RPCA Decomposition

Cited by 48 publications

References 45 publications

Online Reweighted Least Squares Robust PCA

Online Reweighted Least Squares Robust PCA

Integration of Global and Local Features for Specular Reflection Inpainting in Colposcopic Images

Foreground-Background Separation via Generalized Nuclear Norm and Structured Sparse Norm Based Low-Rank and Sparse Decomposition

Contact Info

Product

Resources

About