Related Experiment Video
Updated: Jun 25, 2026
05:33
Three-Dimensional Shape Modeling and Analysis of Brain Structures
Published on: November 14, 2019
7.6K
A Fast and Precise Method to Search for Desirable Rectangular Tumor Regions in Brain MR Images.
Hidenori Takeshima1, Shuki Maruyama2
1Advanced Technology Research Department, Research and Development Center, Canon Medical Systems Corporation, Tokyo, Japan.
Summary
This study introduces a novel method for rapidly and accurately identifying rectangular brain tumor regions in MRI scans. The new approach significantly reduces processing time and improves the precision of tumor localization for better diagnosis.
Area of Science:
- Medical Imaging
- Artificial Intelligence in Medicine
- Neuro-oncology
Background:
- Accurate localization of brain tumors in MRI is crucial for diagnosis and treatment planning.
- Conventional methods for identifying tumor regions can be time-consuming and lack precision.
Purpose of the Study:
- To develop a fast and precise computational method for searching rectangular regions within brain tumor images.
- To enhance the efficiency and accuracy of brain tumor detection using MRI.
Main Methods:
- A novel method combining a segmentation network (EfficientNet-U-Net) and a fast search algorithm utilizing summed-area tables for 3D full search.
- Development of a user-controllable search metric prioritizing isotropic rectangles and higher tumor fractions.
Main Results:
- The proposed method achieved a 100-500 times speedup compared to conventional methods, reducing search time to 8 seconds.
- The new metric resulted in higher tumor fractions and preferred nearly isotropic regions over thin, undesirable ones.
Conclusions:
- The developed method offers a promising solution for fast and precise rectangular tumor region searching in brain MRI.
- This advancement can significantly aid in brain tumor diagnosis and improve the quality of identified tumor regions.
