Robust segmentation of exudates from retinal surface using M-CapsNet via EM routing

Biswal, Birendra; P, Geetha Pavani; T, Prasanna; Karn, Prakash Kumar

doi:10.1016/j.bspc.2021.102770

Cited by 14 publications

(6 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, this method needs of training stage when used in a huge amount of data, which is usually not available in case of EXs overfitting. With some different ideas, B. Biswal et al [14] developed a novel method for segmentation of EXs by using an encoder-decoder style network termed as "deep M-CapsNet" based on Expectation-Maximization (EM) Routing, W. Kusakunniran et al [15] employed the multilayer perceptron techniques is only used to classified initial seeds with high confidences to be EXs regions, C. Huang et al [16] also used superpixel multi-feature extraction and patch-based deep convolutional neural network to coarse segmentation and extraction from digital retinal images. However, this kind of coarse segmentation and extraction are not turned to fine segmentation or do not generalize well.…”

Section: Previous Related Workmentioning

confidence: 99%

Exudates Detection Using Fusion of Histogram-Based Fuzzy C-Means Clustering Based on a New Weight Assignment Scheme

Wisaeng

2023

Preprint

View full text Add to dashboard Cite

Diabetic Retinopathy (DR) is one of the earliest signs of a complication of diabetes mellitus and is also a leading cause of visual blindness globally. Recent studies have proposed that exudates (EXs) are the hallmark of DR severity. The present study aims to accurately and automatically detect EXs related features that are difficult to detect in fundus photography in early stages. The proposed method utilizes a Fusion of Histogram-Based Fuzzy C-Means clustering (FHBFCM) based on a New Weight Assignment Scheme (NWAS), and a set of four selected features from stages of pre-processing in order to evolve the detection method. Through the coarse segmentation stage, the optimal parameter of FHBFCM is trained by the features of DR for detection of EXs diseases through a stepwise enhancement process. The histogram-based is used to find the color intensity in each of the pixels and performed to accomplish RGB color information. This RGB color information is used as the initial cluster centers for creating the appropriate region and generating the homogeneous regions by FCM method. Afterwards, the best expression of NWAS is then used for fine segmentation stage. The experimental results show that the proposed method achieves accuracy values of 96.12%, 97.20%, 93.22% for EXs segmentation on DiaretDB0, DiaretDB1, and STARE fundus image datasets respectively. In conclusion, this study provides a new method for early detection of EXs related features with a competitive accuracy and its ability to outperforms the other stateof-the-art by optimizing the detection quality, and possibly reducing the false positive segmentation in a significant way.

show abstract

Section: Previous Related Workmentioning

confidence: 99%

Exudates Detection Using Fusion of Histogram-Based Fuzzy C-Means Clustering Based on a New Weight Assignment Scheme

Wisaeng

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Some authors, such as Ghosh et al [12] and Syed et al [13] adopted SVM for EXs segmentation. With some different ideas, Biswal et al [14] proposed a new approach for the detection of EXs using an encoder-decoder style network termed "Deep Multitask Capsule Neural Network (M-CapsNet)". The multilayer perception approaches used by Kusakunniran et al [15] are only applied to initial seeds identified as having a high probability of being EXs.…”

Section: Related Workmentioning

confidence: 99%

A Novel Soft Clustering Method for Detection of Exudates

Wisaeng¹

2023

Computer Systems Science and Engineering

View full text Add to dashboard Cite

One of the earliest indications of diabetes consequence is Diabetic Retinopathy (DR), the main contributor to blindness worldwide. Recent studies have proposed that Exudates (EXs) are the hallmark of DR severity. The present study aims to accurately and automatically detect EXs that are difficult to detect in retinal images in the early stages. An improved Fusion of Histogram-Based Fuzzy C-Means Clustering (FHBFCM) by a New Weight Assignment Scheme (NWAS) and a set of four selected features from stages of pre-processing to evolve the detection method is proposed. The features of DR train the optimal parameter of FHBFCM for detecting EXs diseases through a stepwise enhancement method through the coarse segmentation stage. The histogram-based is applied to find the color intensity in each pixel and performed to accomplish Red, Green, and Blue (RGB) color information. This RGB color information is used as the initial cluster centers for creating the appropriate region and generating the homogeneous regions by Fuzzy C-Means (FCM). Afterward, the best expression of NWAS is used for the delicate detection stage. According to the experiment results, the proposed method successfully detects EXs on the retinal image datasets of Dia-retDB0 (Standard Diabetic Retinopathy Database Calibration level 0), DiaretDB1 (Standard Diabetic Retinopathy Database Calibration level 1), and STARE (Structured Analysis of the Retina) with accuracy values of 96.12%, 97.20%, and 93.22%, respectively. As a result, this study proposes a new approach for the early detection of EXs with competitive accuracy and the ability to outperform existing methods by improving the detection quality and perhaps significantly reducing the segmentation of false positives.

show abstract

“…A new CapsNet-based algorithm [21], [22] has recently been proposed, providing viable ideas for further refinement of the results. We believe CapsNet has the potential to achieve better performance by making some changes to the hyperparameters.…”

Section: IImentioning

confidence: 99%

Character Recognition Using Pre-Trained Models and Performance Variants Based on Datasets Size: A Survey

Benaissa¹,

Bahri²,

Allaoui

et al. 2022

ITM Web Conf.

View full text Add to dashboard Cite

The most efficient and beneficial mechanism to the feature of extracting data from an image, has been the Convolutional Neural Network (CNN) and it is used in many fields (Optical character recognition, image classification, object recognition and Facial recognition etc.). In this papier, we studied the character classification problems, using pre-trained models based on Convolutional Neural Network (CNN), and how the performance can change the outcome of dataset that is given. For that, we have used five pre-trained models’ such as VGG16/19, ResNet, Xception et MobileNet. The experiment shows that Xception had the best performance rate compared to other models for all datasets, VGG16/19 performance rate are variants depend on dataset. However, Experiments shows that ResNet achieve the worst accuracy rate compared to other methods.

show abstract

Robust segmentation of exudates from retinal surface using M-CapsNet via EM routing

Cited by 14 publications

References 24 publications

Exudates Detection Using Fusion of Histogram-Based Fuzzy C-Means Clustering Based on a New Weight Assignment Scheme

Exudates Detection Using Fusion of Histogram-Based Fuzzy C-Means Clustering Based on a New Weight Assignment Scheme

A Novel Soft Clustering Method for Detection of Exudates

Character Recognition Using Pre-Trained Models and Performance Variants Based on Datasets Size: A Survey

Contact Info

Product

Resources

About