Article
Computer Science, Artificial Intelligence
Feiping Nie, Jingjing Xue, Danyang Wu, Rong Wang, Hui Li, Xuelong Li
Summary: This paper proposes a solution to solve the k-means problem using coordinate descent method. Through theoretical analysis and experiments, we find that the proposed method performs better in terms of objective function value, local minimum and iterations, and is more robust.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Environmental Sciences
Jingxing Zhu, Feng Wang, Hongjian You
Summary: The existence of multiplicative noise in synthetic aperture radar (SAR) images makes SAR segmentation by fuzzy c-means (FCM) a challenging task. To tackle this problem, we propose two unsupervised FCM segmentation frameworks: LBNL_FCM and GLR_FCM. Both frameworks achieve high segmentation accuracy on simulated and real SAR images.
Article
Computer Science, Artificial Intelligence
Jingwei Chen, Shiyu Xie, Hongyun Jiang, Hui Yang, Feiping Nie
Summary: In this article, we propose k-mRSR, which converts the traditional k-means clustering method into a combinatorial optimization problem. The main advantage of k-mRSR is that it only needs to solve the membership matrix instead of computing the cluster centers in each iteration. Experimental results show that k-mRSR can further decrease (increase) the objective function values of the k-means obtained by Lloyd (CD), while Lloyd (CD) cannot decrease (increase) the objective function obtained by k-mRSR. Furthermore, k-mRSR outperforms both Lloyd and CD in terms of the objective function value and outperforms other state-of-the-art methods in terms of clustering performance.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Sambhav Jain, Reshma Rastogi
Summary: This paper proposes Parametric non-parallel support vector machines for binary pattern classification. The model brings noise resilience and sparsity by intelligently redesigning the Support vector machine optimization. The experimental results validate its scalability for large scale problems.
Article
Computer Science, Information Systems
Wei Chen, Cenyu He, Chunlin Ji, Meiying Zhang, Siyu Chen
Summary: This study presents an improved K-means algorithm for underwater image background segmentation, addressing issues with K value determination and initial centroid position. Experimental results show that the algorithm effectively segments underwater image backgrounds, with low color cast, low contrast, and blurred edges. While the algorithm has higher time cost than existing methods, it proves more efficient than manual segmentation.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Information Systems
Ibrahim Mahmoud El-Henawy, Mostafa Elbaz, Zainab H. Ali, Noha Sakr
Summary: This paper proposes a novel framework that uses 3D MRI images from Kaggle and applies different diverse models to remove Rician and speckle noise for tumor segmentation. Experimental results show the efficiency of the proposed framework against classical filters such as Bilateral, Frost, Kuan, and Lee.
CMC-COMPUTERS MATERIALS & CONTINUA
(2023)
Article
Computer Science, Information Systems
Ali Kavand, Mehdi Bekrani
Summary: This paper proposes a hybrid algorithm using a convolutional neural network (CNN) to reduce speckle noise in medical ultrasonic imaging. Experimental results show that the proposed method outperforms other methods in terms of accuracy and preservation of image details.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Biochemical Research Methods
Jose J. Rico-Jimenez, Dewei Hu, Eric M. Tang, Ipek Oguz, Yuankai K. Tao
Summary: Optical coherence tomography (OCT) is widely used in ophthalmic diagnostic imaging, but image quality varies and can introduce errors in analysis. Frame-averaging is a common method to improve image quality, but it is affected by bulk motion and takes longer acquisition time. A new method called self-fusion, which uses similarity between adjacent frames, is more robust to motion artifacts. In this study, a convolutional neural network was used to perform real-time OCT denoising and offset the computational overhead of self-fusion, resulting in improved image quality.
BIOMEDICAL OPTICS EXPRESS
(2022)
Article
Multidisciplinary Sciences
Yan Hu, Jianfeng Ren, Jianlong Yang, Ruibing Bai, Jiang Liu
Summary: An adaptive denoising algorithm for OCT images is proposed in this paper, which has been proven effective through quantitative and subjective evaluation. The algorithm better removes noise and preserves image details. Furthermore, the proposed algorithm effectively transforms Poisson noise to Gaussian noise so that subsequent Shearlet transform can optimally remove noise.
SCIENTIFIC REPORTS
(2021)
Article
Construction & Building Technology
Qian Huang, Shengyang Chen, Ya Li
Summary: Automatic selection of analysis windows is crucial in conducting accurate HVSR investigations in urban areas. A procedure utilizing K-means cluster analysis is proposed to achieve this goal. Through comparing K-means and hierarchical clustering methods, the characteristics of spectral ratio curves generated by anthropogenic sources (EHVSR) and site information (SHVSR) were examined. An automatic procedure for selecting SHVSR curves using K-means was presented and applied to 24 sites, where the results showed close proximity between SHVSR curve and site transfer function.
CASE STUDIES IN CONSTRUCTION MATERIALS
(2023)
Article
Computer Science, Information Systems
Jingwei Chen, Jianyong Zhu, Bingxia Feng, Shiyu Xie, Hui Yang, Feiping Nie
Summary: In this study, we propose a novel optimization method for the K-Means model that improves the clustering performance by introducing an improved spectral rotation. The experiments demonstrate the effectiveness of the proposed algorithm.
INFORMATION SCIENCES
(2022)
Article
Engineering, Biomedical
Ademola Enitan Ilesanmi, Utairat Chaumrattanakul, Stanislav S. Makhanov
Summary: A new method for semantic segmentation of breast ultrasound images was proposed, using different preprocessing and convolution methods, achieving high scores on malignant and benign breast ultrasound images, and outperforming previous methods.
BIOCYBERNETICS AND BIOMEDICAL ENGINEERING
(2021)
Article
Computer Science, Information Systems
Saleh Naif Almuayqil, Reham Arnous, Noha Sakr, Magdy M. Fadel
Summary: This paper proposes a novel framework for skin segmentation, which consists of two main stages. The first stage aims to remove various types of noises from dermoscopic images, such as hair, speckle, and impulse noise. The second stage focuses on segmenting the dermoscopic images using an attention residual U-shaped network (U-Net).
CMC-COMPUTERS MATERIALS & CONTINUA
(2023)
Article
Computer Science, Information Systems
P. Esther Jebarani, N. Umadevi, Hien Dang, Marc Pomplun
Summary: Breast cancer ranks as the second leading cause of death among women worldwide, making early detection and diagnosis vital for treatment. This research presents a new diagnostic technique by integrating segmentation methods and machine learning, aiming to differentiate between benign and malignant tumors effectively.
Article
Computer Science, Information Systems
Krishna Gopal Dhal, Arunita Das, Buddhadev Sasmal, Swarnajit Ray, Rebika Rai, Arpan Garai
Summary: Image segmentation is a crucial prerequisite for various tasks in digital image processing, and it involves identifying identical segments in an image using well-known clustering techniques. The Fuzzy C-Means algorithm (FCM), being extensively used, has drawbacks such as high computational time complexity, reliance on initial cluster centers and membership matrix, and sensitivity to noise. This paper presents a review of recent literature solutions to overcome these issues and discusses the main challenges in developing improved FCM variants.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)