Edge based method for kidney segmentation in MRI scans
Document Type
Article
Publication Date
1-1-2021
Abstract
The precise and proficient detection of the kidney boundary in low-contrast images is considered as the main difficulty in the detection of kidney boundary in MRI image. The exact identification of a kidney shape in medical images with decreased non-kidney components to acquire insignificant false edge detection is adequately vital for several applications in surgical planning and diagnosis. Low illumination, poor-contrast, image close to the non-uniform state of organs with missing lines, shapes, and edges are considered fundamental difficulties in kidney boundary detection in MRI images. Kidney image edge detection is a significant step in the segmentation procedure because the final appearance and nature of the segmented image depend greatly on the edge detection technique utilized. This study presented a new method of extracting kidney edges from low quality MRI images. The proposed method extracted the unique information of the pixels, which represent the contours of the kidney for segmenting the region. The experimental results on different low-quality kidney MR images showed that the proposed model be able to carry out the effective segmentation of kidney MRI images based on the use of kidney edge components while preserving kidney-segmented edge information from low-contrast MRI images. © 2021, Springer Nature Switzerland AG.
Keywords
Edge-based method, Kidney segmentation, Medical imaging, MRI images
Divisions
Computer
Funders
None
Publication Title
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume
12799
Publisher
Springer Science and Business Media Deutschland GmbH