A Hierarchical Image Matting Model for Blood Vessel Segmentation in Fundus Images

Fan, Zhun; Lu, Jiewei; Wei, Cai-Min; Huang, Han; Cai, Xinye; Chen, Xinjian

doi:10.1109/tip.2018.2885495

Cited by 103 publications

(54 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We gratefully acknowledge DRIU and JL-UNET authors for providing us code/output images; and we faithfully reproduced the PDSN and ML-UNET implementations by confirming that our implementation produces results that are fully consistent with those reported in their papers 8. Note that the AUC measure and precision-recall (PR) curves are not reported for Fan et al[16] -this is also the case in their paper -because their threshold selection strategy is different from the threshold used on the soft output of deep learning methods.…”

supporting

confidence: 78%

“…AUC is considered a particularly important measure for this problem and as Table VIII confirms, MS-DRIS-GP produces the best AUC values on all 3 datasets. Interestingly, amongst the unsupervised methods, the results of Fan et al [16] are comparable to recent deep learning methods. Visual comparisons against the top competing deep learning methods are shown in Fig.…”

Section: Broad Evaluation On a Standard Test-train Configurationmentioning

confidence: 62%

“…It is thus desirable for the representation layer to exhibit robustness against noise components that are particularly similar to vessel structures. The most common way to address this in the literature has been to denoise the image [16], [27], [51], [52] ahead of performing segmentation. Denoising methods however have the limitation of the classical trade-off: while spurious noise like components may be eliminated by denoising algorithms, they also tend to weaken some of the 'true vessel structures'.…”

Section: Data Adaptive Noise Robustnessmentioning

confidence: 99%

See 2 more Smart Citations

Deep Retinal Image Segmentation With Regularization Under Geometric Priors

Cherukuri

Bala

et al. 2020

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

Vessel segmentation of retinal images is a key diagnostic capability in ophthalmology. This problem faces several challenges including low contrast, variable vessel size and thickness, and presence of interfering pathology such as micro-aneurysms and hemorrhages. Early approaches addressing this problem employed hand-crafted filters to capture vessel structures, accompanied by morphological post-processing. More recently, deep learning techniques have been employed with significantly enhanced segmentation accuracy. We propose a novel domain enriched deep network that consists of two components: 1) a representation network that learns geometric features specific to retinal images, and 2) a custom designed computationally efficient residual task network that utilizes the features obtained from the representation layer to perform pixel-level segmentation. The representation and task networks are jointly learned for any given training set. To obtain physically meaningful and practically effective representation filters, we propose two new constraints that are inspired by expected prior structure on these filters: 1) orientation constraint that promotes geometric diversity of curvilinear features, and 2) a data adaptive noise regularizer that penalizes false positives. Multi-scale extensions are developed to enable accurate detection of thin vessels. Experiments performed on three challenging benchmark databases under a variety of training scenarios show that the proposed prior guided deep network outperforms state of the art alternatives as measured by common evaluation metrics, while being more economical in network size and inference time.

show abstract

supporting

confidence: 78%

Section: Broad Evaluation On a Standard Test-train Configurationmentioning

confidence: 62%

Section: Data Adaptive Noise Robustnessmentioning

confidence: 99%

See 1 more Smart Citation

Deep Retinal Image Segmentation With Regularization Under Geometric Priors

Cherukuri

Bala

et al. 2020

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

show abstract

“…Similarly, in order to overcome the problem of isolated points in segmentation, Normalized Cuts (Ncut) is adopted Search new threshold near the original threshold using artificial bee colony algorithm 7Recalculate the value of Ncut under the new threshold (8) Recalculate he value of new cost function (fit ) based on Ncut (9) if the cost function becomes smaller (10) Continue searching new threshold near the original threshold to describe the degree of separation between the two classes [38], which is defined as follows:…”

Section: E Cost Function Construct Based On the Undirected Weightmentioning

confidence: 99%

“…Due to the small difference between the target and the back ground of a complex image, the results of image threshold segmentation are often far from satisfactory. Considering the results of image segmentation are quite different under different thresholds, [7][8][9], providing an accurate, reliable, and effective method for identifying objects in complex background has a wide range of practical applications [10,11]. On the other hand, with the development of computer science and technology, the realtime requirement of image segmentation is improved and finding the exact threshold quickly is also an important part [12,13].…”

Section: Introductionmentioning

confidence: 99%

An Improved Multithreshold Segmentation Algorithm Based on Graph Cuts Applicable for Irregular Image

Wang

et al. 2019

Mathematical Problems in Engineering

View full text Add to dashboard Cite

In order to realize the multithreshold segmentation of images, an improved segmentation algorithm based on graph cut theory using artificial bee colony is proposed. A new weight function based on gray level and the location of pixels is constructed in this paper to calculate the probability that each pixel belongs to the same region. On this basis, a new cost function is reconstructed that can use both square and nonsquare images. Then the optimal threshold of the image is obtained through searching for the minimum value of the cost function using artificial bee colony algorithm. In this paper, public dataset for segmentation and widely used images were measured separately. Experimental results show that the algorithm proposed in this paper can achieve larger Information Entropy (IE), higher Peak Signal to Noise Ratio (PSNR), higher Structural Similarity Index (SSIM), smaller Root Mean Squared Error (RMSE), and shorter time than other image segmentation algorithms.

show abstract

Highly accurate blood vessel segmentation using texture‐based modified K‐means clustering with deep learning model

Lisha¹,

Sulochana

2023

Concurrency and Computation

View full text Add to dashboard Cite

This paper suggests a blood vessel segmentation using the Texture-Based Modified K-means Clustering (TBMKC) technique to lessen the detrimental effects of light lesions. In the suggested method, the K value is chosen automatically by maximizing the local pixel data of the input image. Utilizing the Gabor Filter and Scale-Invariant Feature Transform (SIFT), respectively, texture and statistical characteristics are used to extract the local information. The number of pixels present in each class is used to alter the cluster centroid using Particle Swarm Optimization (PSO) methods. Local adaptive thresholding is then used to separate the blood vessel pattern. The proposed method is invariant to scales and rotations due to SIFT features. The thresholding output is given to the modified DCNN (Deep Convolutional Neural Network). Finally, the performance of the (MDCNN+ TBMKC+ PSO) scheme is computed over other extant schemes. The outcomes of the adopted scheme is determined and tested using the retinal image from DRIVE, STARE, HRF, and CHASE_DBI databases. The accuracy of the proposed method is 13.41%,

show abstract

A Hierarchical Image Matting Model for Blood Vessel Segmentation in Fundus Images

Cited by 103 publications

References 37 publications

Deep Retinal Image Segmentation With Regularization Under Geometric Priors

Deep Retinal Image Segmentation With Regularization Under Geometric Priors

An Improved Multithreshold Segmentation Algorithm Based on Graph Cuts Applicable for Irregular Image

Highly accurate blood vessel segmentation using texture‐based modified K‐means clustering with deep learning model

Contact Info

Product

Resources

About