An improved method of breast MRI segmentation with simplified K-means clustered images

Kang, Donghoon; Shin, Sung Y.; Sung, Chang Oan; Kim, Jung‐Yeul; Pack, Jeong‐Ki; Choi, Hyung Do

doi:10.1145/2103380.2103425

Cited by 11 publications

(5 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By repeating two procedures, namely data clustering and estimating the attributes of a certain tissue class, segmentation and training are carried out simultaneously. The k-means clustering [63], [64], fuzzy c-means clustering [65], [66], Markov random field [67], and expectationmaximization approach [68] are the most often used clustering techniques. This survey focus on deep learning based segmentation methods such as CNN based segmentation and U-net based segmentation.…”

Section: Brain Tumor Segmentationmentioning

confidence: 99%

Comparative Study on Architecture of Deep Neural Networks for Segmentation of Brain Tumor using Magnetic Resonance Images

Preetha,

Priyadarsini,

Nisha

2023

IEEE Access

View full text Add to dashboard Cite

The state-of-the-art works for the segmentation of brain tumor using the images acquired by Magnetic Resonance Imaging (MRI) with their performances are analyzed in this comparative study. First, the architectures of convolutional neural networks (CNN) and the variants of U-shaped Network (U-Net), a kind of Deep Neural Network (DNN) are compared and their differences are highlighted. The publicly available datasets of MRI images specifically Brain Tumor Segmentation (BraTS) are also discussed. Next, the performances of tumor segmentation of various methods in the literature are compared using the parameters such as Dice score and Hausdroff distance ( 95). This study concludes that the U-Net based architectures using the BraTS-2019 dataset outperform well compared with other CNN based architectures.

show abstract

Section: Brain Tumor Segmentationmentioning

confidence: 99%

Comparative Study on Architecture of Deep Neural Networks for Segmentation of Brain Tumor using Magnetic Resonance Images

Preetha,

Priyadarsini,

Nisha

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The strength of K-means method is that it is fully automated while the The wrong choice of k (i.e. the number of clusters) can lead to inaccurate results [15,19]. The strength of Fuzzy C-means is that it minimizes the objective function which results into the optimization of resources deployed for parcellation and the weakness is that it is significantly inefficient if it is used for noisy images and artifacts [15,20].…”

Section: Unsupervised Learning Modelsmentioning

confidence: 99%

Systematic Review of Computational Models for Human Brain Parcellation

Owate

2017

JRRS

View full text Add to dashboard Cite

Introduction: The human brain consists of four main lobar sections: Frontal lobe, Parietal lobe, Temporal lobe and Occipital lobe. Most of the existing models used for the parcellation of brain into these lobes have limited accuracy when applied to ageing brain. Aim: To systematically review the existing models of parcellating brain Magnetic Resonance Images, their strengths and weaknesses, and the possibility of using them for ageing brain. Materials and Methods: PubMed was searched combining search terms for Parcellation, Brain and Magnetic Resonance Imaging (MRI). Articles were considered if they met the following criteria: Parcellation method was indicated, imaging technique was MRI, high resolution anatomical T1-Weighted was used, lobar regions were parcellated, number of lobar regions was indicated. Results: The search resulted into 569 articles. 174 articles (7 from the list of references) were potentially relevant and their abstracts were read. Out of these, 108 were not relevant because they either focused on animal studies, sub-cortical segmentation or tissue segmentation. The full papers of the remaining 66 were reviewed. 39 articles met the inclusion criteria. Various parcellation models were reviewed and summarized into six groups: supervised learning, unsupervised learning, region growing, shape and appearance, energy-based and atlas-based models. Conclusion: All the existing models identified were developed for parcellation of young adult brains and none of them used age-related information. Atlas-based model was found to perform the best among all the models. Future work should consider extending atlas-based model by including ageing information which could make them perform well on ageing brain.

show abstract

“…An unsuitable choice of k can also yield poor and inaccurate results [19]. Global optimum solution in Kmeans algorithm is not guaranteed.…”

Section: K-means Algorithmmentioning

confidence: 99%

Image Segmentation Methods and Applications in MRI Brain Images

Yazdani

Yusof

Karimian

et al. 2015

IETE Technical Review

View full text Add to dashboard Cite

Medical image segmentation plays an important role in medical-imaging applications and they provide a large amount of functional and anatomical information, which improve and facilitate diagnosis and disease therapy planning. However, the existence of image artifacts, such as intensity inhomogeneity, noise and partial volume in magnetic resonance images (MRIs), can adversely affect the quantitative image analysis. There are different segmentation methods in the literature, which segment brain MRI into white matter (WM), gray matter (GM) and cerebrospinal fluid (CSF). However, there is not a common algorithm that can be used for all types of images. We present a critical appraisal of the current status of techniques for MRI segmentation. In this paper, commonly used segmentation algorithms are reviewed and summarized with an emphasis on their characteristics, advantages and disadvantages of these techniques. These are categorized into five different groups based on their workflows and segmentation principles. Different solutions are also proposed to compensate the existing problems in each algorithm. This paper also addresses the issue of quantitative evaluation of segmentation results.

show abstract

An improved method of breast MRI segmentation with simplified K-means clustered images

Cited by 11 publications

References 10 publications

Comparative Study on Architecture of Deep Neural Networks for Segmentation of Brain Tumor using Magnetic Resonance Images

Comparative Study on Architecture of Deep Neural Networks for Segmentation of Brain Tumor using Magnetic Resonance Images

Systematic Review of Computational Models for Human Brain Parcellation

Image Segmentation Methods and Applications in MRI Brain Images

Contact Info

Product

Resources

About