Improved Edge Detection for Brain Tumor Using Multi-threading and Advanced Parallelism

Brain tumors are detected by Computed Tomography (CT) or Magnetic Resonance Imaging (MRI). The medical equipment utilized for MRI scans introduces visible noise into the resulting images. As a result, noise suppression is a critical step in improving the accuracy in picture analysis. There is a lack of standardization in identifying tumors in its early stages, resulting in erroneous conclusions that differ from scan to scan. Various factors contribute to this inaccuracy, especially the noise produced by the MRI scan. Hence, it is essential to use an edge detection mechanism to reduce noise suppression. Brain tumors are identified in a three-step process which includes the pre-processing stage, processing stage, and image segmentation stage. The most important and crucial step in detection is the processing stage which involves the usage of an edge detection algorithm. The edge detection mechanism is implemented on pre-processed images of a brain to detect patterns of brain tumors. The work focuses on the design and implementation of an edge detection algorithm performed in a parallel fashion optimizing the computational efficiency of the edge detector. The time taken to perform edge detection is reduced by employing parallel programming techniques such as MPI and OpenMP.