Brain Tumor Segmentation using U-Net and SegNet

DOI

10.2991/978-94-6463-529-4_18How to use a DOI?

Keywords

U-Net; SegNet; Convolutional Neural Network (CNN); Brain tumor; segmentation; Magnetic Resonance Imaging (MRI)

Abstract

The detection of tumors is the most difficult aspect of quantitative brain tumor evaluation. Magnetic Resonance Imaging (MRI) has gained popularity in recent years due to its non-invasive and powerful soft tissue contrast. MRI is a frequent imaging method used to detect brain cancers. The MRI produces a tremendous amount of data. Heterogeneity, isointensity, and hypointensity are characteristics of tumors that impede manual segmentation in a reasonable amount of time, hence limiting the use of valid quantitative measurements in clinical practice. In clinical practice, manual segmentation tasks are time-intensive and their performance is heavily dependent on the operator’s level of expertise. Also required are accurate and automated tumor segmentation approaches; however, the high spatial and structural variability of brain tumors makes automatic segmentation a challenging task. This paper proposes fully automatic segmentation of brain tumors using convolutional neural networks with encoder-decoders. This work focuses on well-known deep neural networks for semantic segmentation, namely U-Net, and SegNet, for segmenting tumors from Brain MRI data. The networks are trained and evaluated using a publicly available standard dataset, with Dice Similarity Coefficient (DSC) as a metric for the entire predicted image (tumor and background). The average DSC for U-Net on the test dataset is 0.76, while the average DSC for SegNet is 0.67. The examination of results demonstrates that U-Net performs better than SegNet.

Copyright

© 2024 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Pankaj Kasar
AU  - Shivajirao Jadhav
AU  - Vineet Kansal
PY  - 2024
DA  - 2024/10/04
TI  - Brain Tumor Segmentation using U-Net and SegNet
BT  - Proceedings of the International Conference on Signal Processing and Computer Vision (SIPCOV-2023)
PB  - Atlantis Press
SP  - 194
EP  - 206
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6463-529-4_18
DO  - 10.2991/978-94-6463-529-4_18
ID  - Kasar2024
ER  -