Medical imaging techniques are a vital tool in disease diagnosis. The images are being developed to satisfy the growing need for important information from medical image scans by anticipating constitutional tissues for clinical analysis. The application of deep learning techniques is increasing with the demand for automatic diagnosis of medical imaging. Different layers are used in deep learning models to represent data abstraction and construct computational models. Imaging techniques allow medical experts such as radiologists to correctly recognize a patient�s condition, making medical procedures more accessible and automated. The review�s primary goal is to present a study on recent brain tumor detection segmentation and classification approaches. Brain tumors are reviewed because of their importance compared to other tumors and their high illness rate. Many brain tumor segmentation models have been described to grasp these methodologies well, along with their limits and benefits. The study focuses primarily on contemporary deep learning-based brain tumor detection technologies, such as deep generative and deep learning networks. The more advanced and recent techniques available in the literature are also reviewed to describe the methods for performing image segmentation and to emphasize the importance of segmentation models that are not used in real-time due to little or no interaction between clinicians and developers. Most research does not consider the data augmentation element of brain tumor segmentation, which is critical for improving performance. The most challenging feature, or limitation, is the fluctuation in the morphology of tumors or the intensity degree of tumors, both of which still require study in this arena. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.