A Novel Preprocessing Model for Multi Modal Brain MRI image Classification for Stroke Prognosis
Open Access
- 1 February 2025
- journal article
- Published by Binghamton University Libraries in Northeast Journal of Complex Systems (NEJCS)
- Vol. 7, 4
Abstract
Magnetic Resonance Imaging (MRI) is an imaging technique used for the diagnosis and observing the progression in various neurological disorders. Stroke is one of the prominent neurological disorders that creates significant impacts in the patients. It occurs when the blood supply to part of the brain is interrupted or reduced, preventing brain tissues from getting oxygen and nutrients. Multimodal data from various modalities help clinicians in proper prognosis of stroke. Ischemic Stroke Lesion Segmentation Challenge (ISLES22) provides data of stroke data for various stroke patients, the dataset consists of three modalities of data – Fluid Attenuated Inversion Recovery (FLAIR), Apparent Diffusion Coefficient (ADC) and Diffusion-Weighted Imaging (DWI). Multimodal data gives a comprehensive understanding of the brain and the stroke lesions. Complex algorithms and processing steps are required to ensure that the data is prepared for further processing. The objective of this experimental research is to create a novel multimodal preprocessing model that can be used for the preprocessing of the multimodal data from various MRI modalities (FLAIR, DWI and ADC). The proposed model supports the automatic removal of artefacts from the multimodal data, by identifying and applying the best preprocessing techniques for Image Registration (Affine or non-rigid transformations), Normalization (Z Score or min-max normalizations), Denoising Techniques (Gaussian, Median, Non-Local Means, or Anisotropic Diffusion filters) and Bias Filed correction. The best technique is been identified using the evaluation techniques of Dice Coefficient, Signal-to-Noise Ratio (SNR), Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM) and Root Mean Squared Error (RMSE). Preprocessing and preparing of the data are critical for better outcome from the subsequent analysis including segmentation. Here, we propose an Enhanced Image Registration and Artefact Correction (EIRAC) model with Best Image Registration Technique (BIRT) and Multiple Orientation Normalization Denoising and Bias filed correction Parallelly (MONDBP) algorithms for the preprocessing of multimodal MRI images to provides better results for the segmentation of stroke lesions through Machine Learning models.Keywords
This publication has 25 references indexed in Scilit:
- Image normalization techniques and their effect on the robustness and predictive power of breast MRI radiomicsEuropean Journal of Radiology, 2025
- Robust Ensemble of Two Different Multimodal Approaches to Segment 3D Ischemic Stroke Segmentation Using Brain Tumor Representation Among Multiple Center DatasetsJournal of Imaging Informatics in Medicine, 2024
- A feature-enhanced network for stroke lesion segmentation from brain MRI imagesComputers in Biology and Medicine, 2024
- ISLES 2022: A multi-center magnetic resonance imaging stroke lesion segmentation datasetScientific Data, 2022
- Deep convolutional neural network for automatically segmenting acute ischemic stroke lesion in multi-modality MRINeural Computing and Applications, 2019
- Supervised learning based multimodal MRI brain tumour segmentation using texture features from supervoxelsComputer Methods and Programs in Biomedicine, 2018
- V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation2016
- Medical image registration: a reviewComputer Methods in Biomechanics and Biomedical Engineering, 2012
- Adaptive non‐local means denoising of MR images with spatially varying noise levelsJournal of Magnetic Resonance Imaging, 2009
- Measuring Cerebral Blood Flow Using Magnetic Resonance Imaging TechniquesJournal of Cerebral Blood Flow & Metabolism, 1999