Generally, segmentation algorithm is highly task specific. For brain tumor, I recommend the top rank algorithm in BraTs challenge at https://www.med.upenn.edu/cbica/brats2020/rankings.html. The proceedings are here http://www.braintumorsegmentation.org/.
Most of them employed encoder-decoder based U-Net-like architecture, such as nnU-Net or transformer-like network.
The popular techniques used for image segmentation are: thresholding method, edge detection based techniques, region based techniques, clustering based techniques, watershed based techniques, partial differential equation based and artificial neural network based techniques etc.
Thank you Professors Qamar Ul Islam , Zhi-Cheng Li , Khurram Hameed for your very interesting answers and links. This gives a great idea about what should be done.