Image segmentation with complicated background by using seeded region growing

Kang, Chung Chia; Wang, Wen June; Kang, Chung-Hao

doi:10.1016/j.aeue.2012.01.011

Cited by 39 publications

(11 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…After estimating G min and G max using either the K-means or the bi-level methods, the second step consists of implementing a spatial method based on the region growing algorithm to better segment grey levels belonging to the overlap (Tremeau and Borel 1997;Kang, Wang and Kang 2012). Even though this method cannot estimate the size of pores which are below the image resolution, it produces a better segmentation result by taking into account the connectivity between pixels belonging to the pore space.…”

Section: Region Growing Algorithmmentioning

confidence: 99%

Numerical estimation of carbonate rock properties using multiscale images

Jouini

Vega

AlRatrout

2014

Geophysical Prospecting

View full text Add to dashboard Cite

Characterizing the pore space of rock samples using three‐dimensional (3D) X‐ray computed tomography images is a crucial step in digital rock physics. Indeed, the quality of the pore network extracted has a high impact on the prediction of rock properties such as porosity, permeability and elastic moduli. In carbonate rocks, it is usually very difficult to find a single image resolution which fully captures the sample pore network because of the heterogeneities existing at different scales. Hence, to overcome this limitation a multiscale analysis of the pore space may be needed. In this paper, we present a method to estimate porosity and elastic properties of clean carbonate (without clay content) samples from 3D X‐ray microtomography images at multiple resolutions. We perform a three‐phase segmentation to separate grains, pores and unresolved porous phase using 19 μm resolution images of each core plug. Then, we use images with higher resolution (between 0.3 and 2 μm) of microplugs extracted from the core plug samples. These subsets of images are assumed to be representative of the unresolved phase. We estimate the porosity and elastic properties of each sample by extrapolating the microplug properties to the whole unresolved phase. In addition, we compute the absolute permeability using the lattice Boltzmann method on the microplug images due to the low resolution of the core plug images. In order to validate the results of the numerical simulations, we compare our results with available laboratory measurements at the core plug scale. Porosity average simulations for the eight samples agree within 13%. Permeability numerical predictions provide realistic values in the range of experimental data but with a higher relative error. Finally, elastic moduli show the highest disagreements, with simulation error values exceeding 150% for three samples.

show abstract

Section: Region Growing Algorithmmentioning

confidence: 99%

Numerical estimation of carbonate rock properties using multiscale images

Jouini

Vega

AlRatrout

2014

Geophysical Prospecting

View full text Add to dashboard Cite

show abstract

“…In processing medical images, image segmentation plays a pivotal role in segmenting the representation of tissues of interest from the background. As ultrasound (US) images are always crowded with image noises, it becomes practically challenging to segment tumors in BUS images [22,23]. Specifically, image noises that pose challenges to segment BUS images as they effectively include intensity inhomogeneity, low signal-to-noise ratio, and high speckle noise [24].…”

Section: Introductionmentioning

confidence: 99%

A Review on Region of Interest Segmentation Based on Clustering Techniques for Breast Cancer Ultrasound Images

Muhammad

Zeebaree

Brifcani

et al. 2020

JASTT

View full text Add to dashboard Cite

The most prevalent cancer amongst women is woman breast cancer. Ultrasound imaging is a widely employed method for identifying and diagnosing breast abnormalities. Computer-aided diagnosis technologies have lately been developed with ultrasound images to help radiologists enhance the accuracy of the diagnosis. This paper presents several ultrasound image segmentation techniques, mainly focus on eight clustering methods over the last 10 years, and it shows the advantages and disadvantages of these approaches. Breast ultrasound image segmentation is, therefore, still an accessible and challenging issue due to numerous ultrasound artifacts introduced in the imaging process, including high speckle noise, poor contrast, blurry edges, weak signal-to-noise ratio, and intensity inhomogeneity.

show abstract

“…Boundary-based approaches generate transition zones, edges or boundaries in images that separate objects and backgrounds. I Region-based image segmentation is done by grouping pixels with their neighboring pixels which have similar values [12], separating pixels that have different values into other groups, or a combination of both ways [13]. Hybrid based image segmentation can be done by combining two or more approaches in order to produce better segmentation.…”

Section: Introductionmentioning

confidence: 99%

Median Filter For Transition Region Refinement In Image Segmentation

Rosyadi¹,

Suciati²

2018

JUTI

View full text Add to dashboard Cite

2) ABSTRAK Segmentasi citra berbasis transition region adalah salah satu metode segmentasi citra yang sederhana dan efektif. Metode ini mampu mensegmentasi citra yang mengandung objek tunggal atau banyak. Namun, metode ini bergantung pada latar belakang. Metode ini dapat memberikan hasil segmentasi yang buruk jika varians tingkat keabuan pada background cukup tinggi atau background bertekstur. Jadi metode untuk memperbaiki transition region sangatlah diperlukan. Pada penelitian ini diusulkan sebuah metode baru untuk memperbaiki transition region dengan median filter berdasarkan persentase transitional pixels yang saling berdekatan. Transition region diekstraksi dari citra grayscale. Perbaikan pada transition region dilakukan berdasarkan persentase transitional pixels yang saling berdekatan. Beberapa operasi morfologi kemudian dilakukan pada transition region. Setelah itu, proses edge linking digunakan untuk menghubungkan ujung transition region yang putus. Region filling dilakukan untuk mendapatkan area foreground. Akhirnya, citra hasil segmentasi diperoleh dengan menampilkan piksel citra grayscale yang terletak di area foreground. Nilai misclassification error (ME), false negative rate (FNR), dan false positive rate (FPR) dari hasil segmentasi dihitung untuk mengukur kinerja metode yang diusulkan. Kinerja metode yang diusulkan dibandingkan dengan kinerja metode lainnya. Hasil eksperimen menunjukkan bahwa metode yang diusulkan memiliki nilai rata-rata ME, FPR, dan FNR: 0.0297, 0.0209, dan 0.0828 yang menunjukkan bahwa metode yang diusulkan memiliki kinerja yang lebih baik daripada metode-metode lain. Terlebih lagi, metode yang diusulkan bekerja dengan baik pada citra dengan berbagai macam background, terutama pada citra dengan background bertekstur.

show abstract

Image segmentation with complicated background by using seeded region growing

Cited by 39 publications

References 6 publications

Numerical estimation of carbonate rock properties using multiscale images

Numerical estimation of carbonate rock properties using multiscale images

A Review on Region of Interest Segmentation Based on Clustering Techniques for Breast Cancer Ultrasound Images

Median Filter For Transition Region Refinement In Image Segmentation

Contact Info

Product

Resources

About