What problem does this paper attempt to address?

The problem that this paper attempts to solve is the automatic segmentation of vascular structures in retinal images. Specifically, the authors propose a method based on two - dimensional Morlet wavelet transform and supervised classification to achieve the automatic segmentation of blood vessels in retinal images. ### Problem Background The evaluation of retinal blood vessels is of great significance for diagnosing a variety of vascular and non - vascular diseases (such as hypertension, diabetes, atherosclerosis, cardiovascular diseases, and stroke, etc.). In particular, diabetic retinopathy, due to its changes in vascular structure and distribution (such as neovascularization), is one of the main causes of blindness in adults. Early identification of these changes can prevent severe vision loss. Therefore, developing an effective automatic blood vessel segmentation method is crucial for early diagnosis and treatment. ### Limitations of Existing Methods Although many blood vessel segmentation methods have been proposed, visual inspection and receiver operating characteristic (ROC) analysis show that there is still room for improvement in existing methods. In addition, many existing algorithms rely on the configuration of multiple parameters, which makes it difficult for untrained community health workers to use these techniques. Therefore, a framework that only depends on manually segmented images and is easy to use is required. ### Solutions Proposed in the Paper To solve the above problems, this paper proposes a new method that uses two - dimensional Morlet wavelet transform to extract pixel features and combines supervised classification for blood vessel segmentation. The specific steps are as follows: 1. **Feature Extraction**: - Use two - dimensional Morlet wavelet transform to calculate the response of each pixel at multiple scales. - The feature vector consists of pixel intensity and its continuous two - dimensional Morlet wavelet transform responses at different scales. 2. **Classifier Selection**: - Use a Bayesian classifier, in which the class - conditional probability density function (likelihood function) is described as a Gaussian mixture model (GMM) to achieve rapid classification and model complex decision boundaries. - At the same time, a performance comparison with the linear minimum mean - square - error (LMSE) classifier is carried out. 3. **Training and Evaluation**: - Build a training set based on manually segmented training images to ensure that the classifier can learn the correct features. - Conduct performance evaluation on the publicly available DRIVE and STARE databases, and use the ROC curve to measure the classification effect. ### Main Contributions - Propose a blood vessel segmentation method based on two - dimensional Morlet wavelet transform and supervised classification. - Achieve the modeling of complex decision boundaries through the Gaussian mixture model, improving the classification accuracy. - Obtain an area under the ROC curve (AUC) of 0.9598 on the DRIVE database, which is slightly better than the method proposed by Staal et al. Through this method, the paper provides an effective and easy - to - use automatic retinal blood vessel segmentation tool, which is helpful for the early diagnosis and treatment of various eye diseases.

Neutrino Physics and Astrophysics : Highlights

Retinal vessel segmentation based on Fully Convolutional Neural Networks

Accurate Retinal Vessel Segmentation via Octave Convolution Neural Network

Automated Retinal Vessel Segmentation Based on Morphological Preprocessing and 2D-Gabor Wavelets

An improved semi-supervised segmentation of the retinal vasculature using curvelet-based contrast adjustment and generalized linear model

Simultaneous segmentation and classification of the retinal arteries and veins from color fundus images

An accurate unsupervised extraction of retinal vasculature using curvelet transform and classical morphological operators

Software multiplataforma para a segmentação de vasos sanguíneos em imagens da retina

Retinal blood vessel segmentation in high resolution fundus photographs using automated feature parameter estimation

Adaptive Feature Enhancement of Retinal Vascular Images Using Wavelet-Based Multiresolution Analysis

Retinal Blood Vessel Segmentation Using Hybrid Features and Multi-Layer Perceptron Neural Networks

Microscopic retinal blood vessels detection and segmentation using support vector machine and K-nearest neighbors

Retina blood vessels segmentation based on the combination of the supervised and unsupervised methods

FCM Based Blood Vessel Segmentation Method for Retinal Images

Retinal vessel segmentation by probing adaptive to lighting variations

Advanced OCTA imaging segmentation: Unsupervised, non-linear retinal vessel detection using modified self-organizing maps and joint MGRF modeling

Image decomposition based segmentation of retinal vessels

Multiscale Joint Optimization Strategy for Retinal Vascular Segmentation

Fundus Image Based Retinal Vessel Segmentation Utilizing a Fast and Accurate Fully Convolutional Network

Based on Retinal Vessel Model to Segment the Image and to Detect the Vessels

Retinal Vessel Segmentation Using Multi-Scale Residual Convolutional Neural Network (MSR-Net) Combined with Generative Adversarial Networks