Neural networks for medical image segmentation

Radjabov, S. S.; Khashimov, Akhmad; Urinov, Elmurod; Kakharov, Shukrullo

doi:10.1109/icisct52966.2021.9670147

Search citation statements

Order By: Relevance

Paper Sections

Select...

Citation Types

Supporting

Mentioning

Contrasting

Year Published

2024

Publication Types

Select...

Other3

Article1

Relationship

Self Cite0

Independent4

Authors

Journals

Cited by 4 publications

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

Image segmentation algorithms

Mirzaev,

Radjabov,

Khashimov

et al. 2024

BIO Web Conf.

View full text Add to dashboard Cite

This article focuses on the methods of segmentation of kidney images, with the main emphasis on segmentation methods based on neural networks. During this work, we got acquainted with neural network-based algorithms and decided to use the U-Net algorithm for segmentation. A neural network architecture mainly consists of a descending (left) and an expanding (right) part. The structure of U-Net neural network architectures, which are currently widely used for medical images, have been investigated and experimental studies have been carried out on extracting the kidney region in medical images. This paper uses the Unet neural network to segment kidney images and further refines its results using morphological operations. In the next step, the kidney area itself was extracted from the main abdominal image. A dataset used in the Kidney Tumor Segmentation Challenge (KiTS19) was used to train the neural network, where kidney regions were defined in the images. As an example, we took medical images. But in general, it will be possible to extract segments from images of energy objects using the proposed segmentation algorithms.

show abstract