Assist. Lec. Noor Sabah from the College of Biomedical Informatics participated in publishing a scientific paper as part of the proceedings of an international peer-reviewed conference. The paper was published as a chapter in a book released by the global publisher Springer Nature, which is indexed in the Scopus database, within a specialized scientific series in the fields of artificial intelligence and computer science.

The paper is titled:
Brain Tumor Detection Based on Deep Learning and MRI
It addresses one of the most critical health challenges related to the diagnosis of brain tumors, which requires high accuracy and speed to support effective treatment planning.

The research presents an intelligent system based on deep learning techniques for detecting and classifying brain tumors using three-dimensional FLAIR MRI images. A 3D U-Net architecture was proposed, achieving a high accuracy of 0.976, reflecting the system’s efficiency in accurately identifying tumors.

The system also extracts important quantitative features of the tumor, including size, orientation, axis lengths, and spatial location, which are essential indicators in supporting clinical medical decision-making.

The model was trained and validated using a large public dataset comprising 484 patients, enhancing the reliability of the results. The system features an easy-to-use interface and real-time processing capabilities, facilitating its integration into clinical environments.

The contributions of this research lie in achieving high performance in the real-time three-dimensional detection and segmentation of brain tumors, along with an accurate representation of tumor shape and size, which aids in improving treatment planning, monitoring disease progression, and enhancing patient outcomes. The study also suggests future prospects for development through the integration of multimodal imaging, transfer learning techniques, and longitudinal studies.

To view the research:
https://link.springer.com/chapter/10.1007/978-3-032-00232-7_35