This notebook contains the implementation of a neural network U-net for brain tumor segmentation in magnetic resonance images. The images and masks used in the training are available at this link: https://www.kaggle.com/datasets/mateuszbuda/lgg-mri-segmentation
- Problem definition
Brain tumors represent a complex and serious condition that affects millions of individuals worldwide. The diagnosis of these tumors can be a challenging process, especially in cases of small or diffuse tumors. Imaging technology, such as magnetic resonance imaging (MRI), is an important tool for aiding in the diagnosis of brain tumors. However, image analysis can be time-consuming and requires specialized knowledge.
- A possible solution
In this context, the use of artificial intelligence (AI) algorithms emerges as a promising alternative to assist in the identification of brain tumors in MRI images. AI algorithms can be trained to identify specific patterns and features in medical data, aiding in the detection of tumors with greater accuracy and efficiency. Additionally, AI can be used to segment the tumors and classify them according to their degree of malignancy.
The U-net architecture is a convolutional neural network that has been widely used in medical image segmentation tasks. Initially proposed for the segmentation of biological cell images, its application was extended to the segmentation of brain tumors in MRI images.
- Some results
- References
-
RONNEBERGER, O.; FISCHER, P.; BROX, T. U-Net: Convolutional Networks for Biomedical Image Segmentation. arXiv preprint arXiv:1505.04597, 2015.
-
Zheng, P., Zhu, X. & Guo, W. Brain tumour segmentation based on an improved U-Net. BMC Med Imaging 22, 199 (2022).