A New Supervised Method for Blood Vessel Segmentation in Retinal Images by Using Gray-Level and Moment Invariants-Based Features

By Diego Marín, Arturo Aquino, Manuel Emilio Gegúndez-Arias, José Manuel Bravo. IEEE Transactions on Medical Imaging, 2011

1,008 citations recorded by the source corpus.

This paper presents a new supervised method for blood vessel detection in digital retinal images. This method uses a neural network (NN) scheme for pixel classification and computes a 7-D vector composed of gray-level and moment invariants-based features for pixel representation. The method was evaluated on the publicly available DRIVE and STARE databases, widely used for this purpose, since they contain retinal images where the vascular structure has been precisely marked by experts. Method performance on both sets of test images is better than other existing solutions in literature. The method proves especially accurate for vessel detection in STARE images. Its application to this database (even when the NN was trained on the DRIVE database) outperforms all analyzed segmentation approaches. Its effectiveness and robustness with different image conditions, together with its simplicity and fast implementation, make this blood vessel segmentation proposal suitable for retinal image computer analyses such as automated screening for early diabetic retinopathy detection.

Subfields: Computer Vision and Pattern Recognition, Ophthalmology, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/1c98f7a90ac165bb8be69732c1b141a0774c7ebc

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Kwang-Ting Cheng
2. Xin Yang
3. Zengqiang Yan
4. Huan Fu
5. Gary G. Yen
6. Stephen Lin
7. Yao Zhou
8. Zhang Yi

In Ophthalmology

1. Kwang-Ting Cheng
2. Xin Yang
3. Zengqiang Yan
4. Huan Fu
5. Gary G. Yen
6. Yao Zhou
7. Zhang Yi
8. Jiang Liu

In Radiology, Nuclear Medicine and Imaging

1. Kwang-Ting Cheng
2. Xin Yang
3. Zengqiang Yan
4. Huan Fu
5. Gary G. Yen
6. Yao Zhou
7. Zhang Yi
8. Jiang Liu

Citations in the corpus

• Joint Segment-Level and Pixel-Wise Losses for Deep Learning Based Retinal Vessel Segmentation (2018, methodology): “[21] trained a neural network on a 7-D vector feature space including gray-level and moment invariants-based features.”
• A Cross-Modality Learning Approach for Vessel Segmentation in Retinal Images (2016, methodology): “Compared to other supervised methods [8], [16], [17], [24], our method can achieve unsupervised learning of the optimal feature for vessel detection.”
• Segmenting Retinal Blood Vessels With Deep Neural Networks (2016, neutral): “Among them, [18] attained particularly good AUC.”
• Blood Vessel Segmentation of Fundus Images by Major Vessel Extraction and Subimage Classification (2015, methodology): “…¯ v) q T W s ( x , y ) (u, v) (16) η pq (s) = μ pq (s) μ 00 (s) γ , γ = p + q 2 + 1 = 2 (17) θ 1 (s) = η 20 (s) + η 02 (s), f 6 (s) = |log(θ 1 (s)))| The seven features f 1 (s) to f 7 (s) are evaluated for each vessel subimage pixel by varying window side-lengths s = [3] [5] [7] [9] [11] [13] [15].”
• An effective retinal blood vessel segmentation method using multi-scale line detection (2013, methodology): “Recently, there is an emerging trend of using supervised methods [6–8,21–26] to perform this segmentation problem.”
• Retinal vessel segmentation by improved matched filtering: evaluation on a new high-resolution fundus image database (2013, methodology): “[20] developed a supervised classi fi cation-based technique, which used a seven-dimensional feature vector composed of grey-level and invariant moment-based features.”
• DUNet: A deformable network for retinal vessel segmentation (2018, background): “D vector composed of gray-level and moment invariants-based features for pixel representation [25].”
• Automated Vessel Segmentation Using Infinite Perimeter Active Contour Model with Hybrid Region Information with Application to Retinal Images (2015, methodology): “003 lower than the method proposed by [21].”
• A Discriminatively Trained Fully Connected Conditional Random Field Model for Blood Vessel Segmentation in Fundus Images (2017, methodology): “FOV masks are not provided in the original set, so the masks built by Mar´ın et al. 3 [10] were used.”
• DeepVessel: Retinal Vessel Segmentation via Deep Learning and Conditional Random Field (2016, methodology): “employed the gray-level vector and moment invariant features to classify each pixel using a neural network [9].”
• Hierarchical retinal blood vessel segmentation based on feature and ensemble learning (2015, methodology): “However, fine vessels are not important for clinical importance [20,23].”
• Blood vessel segmentation methodologies in retinal images - A survey (2012, methodology): “The background homogenization followed by the use of gray level and momentinvariant based features for supervised classification [39] is observed to be training set robust.”
• Evolutionary Compression of Deep Neural Networks for Biomedical Image Segmentation (2020, methodology): “For retinal vessel segmentation, in [33], a neural network with three hidden layers is employed as a pixelwise classifier to label a pixel as vessel or nonvessel, where a set of hand-crafted features is manually extracted as input to the neural network.”
• Computer Methods and Programs in Biomedicine (neutral): “Neural networks are used in [64] to segment retinal vessels.”
• An Ensemble Classification-Based Approach Applied to Retinal Blood Vessel Segmentation (2012, background): “case of the cross-trained classifier for DRIVE and STARE; the cross-trained accuracy of Marin [24] is slightly better.”