Methods of Brain Extraction from Magnetic Resonance Images of Human Head: A Review.

Medical images are providing vital information to aid physicians in diagnosing a disease afflicting the organ of a human body. Magnetic resonance imaging is an important imaging modality in capturing the soft tissues of the brain. Segmenting and extracting the brain is essential in studying the structure and pathological condition of brain. There are several methods that are developed for this purpose. Researchers in brain extraction or segmentation need to know the current status of the work that have been done. Such an information is also important for improving the existing method to get more accurate results or to reduce the complexity of the algorithm. In this paper we review the classical methods and convolutional neural network-based deep learning brain extraction methods.

Isolation, characterization and plant growth-promoting effects of sorghum [Sorghum bicolor (L.) moench] root-associated rhizobacteria and their potential role in drought mitigation.

Drought is a major constraint throughout the world, and it creates a major yield loss by changing the plant metabolic process. However, the negative effects of drought on plant growth and development were alleviated by using plant growth-promoting bacteria. With these backgrounds, the study was conducted to identify the drought-tolerant endophytic bacteria and to know their plant growth promotion (PGP) effect on sorghum plants under drought conditions. From sorghum root, Acinetobacter pittii, Bacillus lichiniformis, Bacillus sp., Pseudacidovorax intermedius, and Acinetobacter baumannii strains were isolated and identified through 16S rRNA sequencing. These strains had higher levels of proline, protein, exopolysaccharides (EPS), 1-aminocyclopropane-l-carboxylic acid (ACC) deaminase, indole-3-Acetic Acid (IAA), and gibberellic acid (GA). An experiment was carried out in the laboratory to evaluate the effects of three drought-tolerant strains, A. pittii, Bacillus sp., and P. intermedius, on the growth of sorghum seedlings. Whereas root length (RL), shoot length (SL), seedling vigor index (SVI), and total dry matter production (TDM) were more in the Bacillus sp., and P. intermedius inoculated plants in both stress and non-stress condition. Principle component analysis revealed that Bacillus sp. and P. intermedius improved the growth characteristics and protect the seedling from water stress situations. A correlation study between the variables showed a positive significant correlation between all variables except root: shoot ratio (RSR) and SL. Variable RSR was not significantly correlated with GP, GRI, and SL; SVI and TDM showed a non-significant correlation with RSR.

Three-Phase Automatic Brain Tumor Diagnosis System Using Patches Based Updated Run Length Region Growing Technique.

Manually finding and segmenting brain tumor is a tedious process in MR brain images due to the unpredictable appearance of tissues with a different pattern, contour, mass, and positions. The proposed work has three phases automatic tumor diagnosis system for tumorous slice detection, segmentation, and visualization from MRI human head volumes. The proposed method has an automatic classification followed by segmentation and is called as patch-based updated run length region growing technique (PR2G). In the first phase, classification is done through training and testing process using SVM classifier with 8 × 8 patches. Three optimal features are chosen using infinite feature selection (IFS) method. The purpose of the first phase is to automatically cluster the input MRI image into a normal or tumorous slice and localize the tumor. The second phase aims to segment the tumor in abnormal tumorous slices identified by the first phase using run length region growing technique. Finally, the third phase contains a post metric evaluation like 3D tumor volume construction and estimation from actual and segmented tumor volume using Carelieri's estimator. Classification accuracy is measured using sensitivity, specificity, accuracy, and error rates also calculated using false alarm (FA) and missed alarm (MA). Segmentation accuracy is calculated using Dice similarity, positive predictive value (PPV), sensitivity, and accuracy. Datasets used for this experiment are collected from whole brain atlas (WBA) and BraTS repositories. Experimental results show that the PR2G achieves 97% of classification accuracy and 80% of Dice segmentation accuracy.

Automatic brain extraction methods for T1 magnetic resonance images using region labeling and morphological operations.

In this work we propose two brain extraction methods (BEM) that solely depend on the brain anatomy and its intensity characteristics. Our methods are simple, unsupervised and knowledge based. Using an adaptive intensity thresholding method on the magnetic resonance images of head scans, a binary image is obtained. The binary image is labeled using the anatomical facts that the scalp is the boundary between head and background, and the skull is the boundary separating brain and scalp. A run length scheme is applied on the labeled image to get a rough brain mask. Morphological operations are then performed to obtain the fine brain on the assumption that brain is the largest connected component (LCC). But the LCC concept failed to work on some slices where brain is composed of more than one connected component. To solve this problem a 3-D approach is introduced in the BEM. Experimental results on 61 sets of T1 scans taken from MRI scan center and neuroimage web services showed that our methods give better results than the popular methods, FSL's Brain Extraction Tool (BET), BrainSuite's Brain Surface Extractor (BSE) gives results comparable to that of Model-based Level Sets (MLS) and works well even where MLS failed. The average Dice similarity index computed using the "Gold standard" and the specificity values are 0.938 and 0.992, respectively, which are higher than that for BET, BSE and MLS. The average processing time by one of our methods is ≈1s/slice, which is smaller than for MLS, which is ≈4s/slice. One of our methods produces the lowest false positive rate of 0.075, which is smaller than that for BSE, BET and MLS. It is independent of imaging orientation and works well for slices with abnormal features like tumor and lesion in which the existing methods fail in certain cases.

Fully automatic brain extraction algorithm for axial T2-weighted magnetic resonance images.

In this paper we propose two brain extraction algorithms (BEA) for T2-weighted magnetic resonance imaging (MRI) scans. The T2-weighted image is first filtered with a low pass filter (LPF) to remove or subdue the background noise. Then the image is diffused to enhance the brain boundaries. Using Ridler's method a threshold value for intensity is obtained. Using the threshold value a rough binary brain image is obtained. By performing morphological operations and using the largest connected component (LCC) analysis, a brain mask is obtained from which the brain is extracted. This method uses only 2D information of slices and is named as 2D-BEA. The concept of LCC failed in few slices. To overcome this problem, 3D information available in adjacent slices is used which resulted in 3D-BEA. Experimental results on 20 MRI data sets show that the proposed 3D-BEA gave excellent results. The performance of this 3D-BEA is better than 2D-BEA and other popular methods, brain extraction tool (BET) and brain surface extractor (BSE).