3D brain tumor segmentation in MRI images based on a modified PSO technique

Three-dimensional (3D) brain tumor segmentation is a clinical requirement for brain tumor diagnosis and radiotherapy planning. This is a challenging task due to variation in type, size, location, and shape of tumors. Several methods such as particle swarm optimization (PSO) algorithm formed a topological relationship for the slices that converts 2D images into 3D magnetic resonance imaging (MRI) images which does not provide accurate results and they depend on the number of input sections, positions, and the shape of the MRI images. In this article, we propose an efficient 3D brain tumor segmentation technique called modified particle swarm optimization. Also, segmentation results are compared with Darwinian particle swarm optimization (DPSO) and fractional-order Darwinian particle swarm optimization (FODPSO) approaches. The experimental results show that our method succeeded 3D segmentation with 97.6% of accuracy rate more efficient if compared with the DPSO and FODPSO methods with 78.1% and 70.21% for the case of T1-C modality.     

Deep learning and machine learning-based early survival predictions of glioblastoma patients using pre-operative three-dimensional brain magnetic resonance imaging modalities

To propose and implement an automated machine learning (ML) based methodology to predict the overall survival of glioblastoma multiforme (GBM) patients. In the proposed methodology, we used deep learning (DL) based 3D U-shaped Convolutional Neural Network inspired encoder-decoder architecture to segment the brain tumor. Further, feature extraction was performed on these segmented and raw magnetic resonance imaging (MRI) scans using a pre-trained 2D residual neural network. The dimension-reduced principal components were integrated with clinical data and the handcrafted features of tumor subregions to compare the performance of regression-based automated ML techniques. Through the proposed methodology, we achieved the mean squared error (MSE) of 87 067.328, median squared error of 30 915.66, and a SpearmanR correlation of 0.326 for survival prediction (SP) with the validation set of Multimodal Brain Tumor Segmentation 2020 dataset. These results made the MSE far better than the existing automated techniques for the same patients. Automated SP of GBM patients is a crucial topic with its relevance in clinical use. The results proved that DL-based feature extraction using 2D pre-trained networks is better than many heavily trained 3D and 2D prediction models from scratch. The ensembled approach has produced better results than single models. The most crucial feature affecting GBM patients' survival is the patient's age, as per the feature importance plots presented in this work. The most critical MRI modality for SP of GBM patients is the T2 fluid attenuated inversion recovery, as evident from the feature importance plots.     

Unsupervised tissue classification of brain MR images for voxel‐based morphometry analysis

In this article, a fully unsupervised method for brain tissue segmentation of T1‐weighted MRI 3D volumes is proposed. The method uses the Fuzzy C‐Means (FCM) clustering algorithm and a Fully Connected Cascade Neural Network (FCCNN) classifier. Traditional manual segmentation methods require neuro‐radiological expertise and significant time while semiautomatic methods depend on parameter's setup and trial‐and‐error methodologies that may lead to high intraoperator/interoperator variability. The proposed method selects the most useful MRI data according to FCM fuzziness values and trains the FCCNN to learn to classify brain’ tissues into White Matter, Gray Matter, and Cerebro‐Spinal Fluid in an unsupervised way. The method has been tested on the IBSR dataset, on the BrainWeb Phantom, on the BrainWeb SBD dataset, and on the real dataset “University of Palermo Policlinico Hospital” (UPPH), Italy. Sensitivity, Specificity, Dice and F‐Factor scores have been calculated on the IBSR and BrainWeb datasets segmented using the proposed method, the FCM algorithm, and two state‐of‐the‐art brain segmentation software packages (FSL and SPM) to prove the effectiveness of the proposed approach. A qualitative evaluation involving a group of five expert radiologists has been performed segmenting the real dataset using the proposed approach and the comparison algorithms. Finally, a usability analysis on the proposed method and reference methods has been carried out from the same group of expert radiologists. The achieved results show that the segmentations of the proposed method are comparable or better than the reference methods with a better usability and degree of acceptance.     

Using a convolutional neural network model to derive imaging landmarks for lumbar spine numbering on axial magnetic resonance images

Understanding the axial lumbar spine anatomy, including knowledge of the relationship between the lumbar spine level and other paraspinal structures, is important for diagnosing and treating diseases. The purpose of this study was to validate the accuracy of a convolutional neural network (CNN) model in lumbar spine level numbering on axial magnetic resonance (MR) images and to find the appropriate anatomic landmarks for numbering using a class activation map (CAM). A total of 6055 axial MR images of the lumbar spine from the L1-2 to L5-S1 disc levels were obtained to train and validate the CNN model. MR images were acquired using three 3-Tesla machines. The algorithm was developed with three models, and the best-performing model was selected. The external validation set (n = 493) was obtained from other institutions using various machines. The accuracy of the numbering was analyzed using a confusion matrix and receiver operating characteristic curves. The CAMs were reviewed, and the identified anatomic structures were investigated. A reader study was performed by three radiologists, and their accuracy was compared with that of the model. The overall accuracy of the best-performing model for lumbar spine numbering was 0.98 on internal validation and 0.95 on external validation. For the CAM review, mappings concentrated on both paraspinal areas, including the kidney, back muscles, and ilium according to the level. Top-1 and top-2 accuracies of the reviewers ranged between 0.56–0.75, and 0.84–0.93, respectively. After reviewing the CAMs, the accuracy increased to 0.75–0.78 and 0.93–0.98, respectively. A CNN model can accurately determine the level of the lumbar spine on axial MR images, and the configuration of muscles can be used to determine the lumbar level.     

Efficient image compression techniques for compressing multimodal medical images using neural network radial basis function approach

Image compression technique is used to reduce the number of bits required in representing image, which helps to reduce the storage space and transmission cost. Image compression techniques are widely used in many applications especially, medical field. Large amount of medical image sequences are available in various hospitals and medical organizations. Large images can be compressed into smaller size images, so that the memory occupation of the image is considerably reduced. Image compression techniques are used to reduce the number of pixels in the input image, which is also used to reduce the broadcast and transmission cost in efficient form. This is capable by compressing different types of medical images giving better compression ratio (CR), low mean square error (MSE), bits per pixel (BPP), high peak signal to noise ratio (PSNR), input image memory size and size of the compressed image, minimum memory requirement and computational time. The pixels and the other contents of the images are less variant during the compression process. This work outlines the different compression methods such as Huffman, fractal, neural network back propagation (NNBP) and neural network radial basis function (NNRBF) applied to medical images such as MR and CT images. Experimental results show that the NNRBF technique achieves a higher CR, BPP and PSNR, with less MSE on CT and MR images when compared with Huffman, fractal and NNBP techniques.     

Dynamic multi-swarm particle swarm optimizer using parallel PC cluster systems for global optimization of large-scale multimodal functions

This article presents a novel parallel multi-swarm optimization (PMSO) algorithm with the aim of enhancing the search ability of standard single-swarm PSOs for global optimization of very large-scale multimodal functions. Different from the existing multi-swarm structures, the multiple swarms work in parallel, and the search space is partitioned evenly and dynamically assigned in a weighted manner via the roulette wheel selection (RWS) mechanism. This parallel, distributed framework of the PMSO algorithm is developed based on a master–slave paradigm, which is implemented on a cluster of PCs using message passing interface (MPI) for information interchange among swarms. The PMSO algorithm handles multiple swarms simultaneously and each swarm performs PSO operations of its own independently. In particular, one swarm is designated for global search and the others are for local search. The first part of the experimental comparison is made among the PMSO, standard PSO, and two state-of-the-art algorithms (CTSS and CLPSO) in terms of various un-rotated and rotated benchmark functions taken from the literature. In the second part, the proposed multi-swarm algorithm is tested on large-scale multimodal benchmark functions up to 300 dimensions. The results of the PMSO algorithm show great promise in solving high-dimensional problems.     

Edge detection using Chebyshev's orthogonal polynomial and brain extraction from magnetic resonance images of human head

In this article, we propose a new edge detecting method based on the transform coefficients obtained by a point spread function constructed out of Chebyshev's orthogonal polynomials. This edge detector finds edges similar to that of Prewitt and Roberts but is robust against additive and multiplicative noises. We also propose a new scheme to extract brain portion from the magnetic resonance images (MRI) of human head scan by making use of the of the new edge detector. The proposed scheme involves edge detection, morphological operations, and largest connected component analysis. Experiments conducted by applying the proposed scheme on 19 volumes of MRI collected from Internet Brain Segmentation Repository (IBSR) show that the proposed brain extraction scheme performed better than the popular Brain Extraction Tool (BET). The performance of the proposed scheme is measured by computing the Dice coefficient (D) and Jaccard similarity index (J). The proposed method produced a value of 0.9068 for D and 0.8321 for J.     

Multimodal magnetic resonance imaging for Alzheimer's disease diagnosis using hybrid features extraction and ensemble support vector machines

Magnetic resonance imaging (MRI) is increasingly used in the diagnosis of Alzheimer's disease (AD) in order to identify abnormalities in the brain. Indeed, cortical atrophy, a powerful biomarker for AD, can be detected using structural MRI (sMRI), but it cannot detect impairment in the integrity of the white matter (WM) preceding cortical atrophy. The early detection of these changes is made possible by the novel MRI modality known as diffusion tensor imaging (DTI). In this study, we integrate DTI and sMRI as complementary imaging modalities for the early detection of AD in order to create an effective computer-assisted diagnosis tool. The fused Bag-of-Features (BoF) with Speeded-Up Robust Features (SURF) and modified AlexNet convolutional neural network (CNN) are utilized to extract local and deep features. This is applied to DTI scalar metrics (fractional anisotropy and diffusivity metric) and segmented gray matter images from T1-weighted MRI images. Then, the classification of local unimodal and deep multimodal features is first performed using support vector machine (SVM) classifiers. Then, the majority voting technique is adopted to predict the final decision from the ensemble SVMs. The study is directed toward the classification of AD versus mild cognitive impairment (MCI) versus cognitively normal (CN) subjects. Our proposed method achieved an accuracy of 98.42% and demonstrated the robustness of multimodality imaging fusion.     

A fast and robust segmentation of magnetic resonance brain images using a combination of the pyramidal approach and level set method

We propose in this article an approach to optimize the processing time and to improve the quality of brain magnetic resonance images segmentation. Level set method (LSM) was adopted with a periodic reinitialization process to prevent the LS function from being too steep or too flat near the interface. Although it is used to maintain the stability of the interface evolution and gives interesting results, it requires a longer processing time. To overcome this disadvantage and reduce the processing time, we propose a hybridization with a regular Gaussian pyramid, which reduces the resolution of the initial image and prevents the possibility of local minima. To compare the different segmentation algorithms, we used six types of quality measurements: specificity, sensitivity, Dice similarity, the Jaccard index, and the correctly and incorrectly marked pixels. A comparison between the results obtained by LSM, LSM with reinitialization, the approach of Barman et al., An International Journal 1 (2011), particle swarm optimization based on the Chan and Vese model (Mandal et al., Engineering Applications of Artificial Intelligence 35 (2014), 199‐214) and by our hybrid approach reveals a clear efficiency of our hybridization strategy. The processing time was significantly reduced, and the quality of segmentation was improved. © 2016 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 26, 243–253, 2016     

A novel Markov random field model based on region adjacency graph for T1 magnetic resonance imaging brain segmentation

Tissue segmentation in magnetic resonance brain scans is the most critical task in different aspects of brain analysis. Because manual segmentation of brain magnetic resonance imaging (MRI) images is a time‐consuming and labor‐intensive procedure, automatic image segmentation is widely used for this purpose. As Markov Random Field (MRF) model provides a powerful tool for segmentation of images with a high level of artifacts, it has been considered as a superior method. But because of the high computational cost of MRF, it is not appropriate for online processing. This article has proposed a novel method based on a proper combination of MRF model and watershed algorithm in order to alleviate the MRF's drawbacks. Results illustrate that the proposed method has a good ability in MRI image segmentation, and also decreases the computational time effectively, which is a valuable improvement in the online applications. © 2017 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 27, 78–88, 2017     

A modified threshold score-based multilevel thresholding segmentation technique for brain magnetic resonance images using opposition-based learning hybrid rice optimization algorithm

In modern clinical diagnostics, magnetic resonance imaging (MRI) is frequently used for brain tumor detection because of its high resolution of soft tissues, which plays a crucial role in the prevention, detection, and treatment planning. Therefore, it is meaningful to obtain high-quality MR images by automatic thresholding for aiding diagnosis. Most multilevel thresholding techniques are based on histograms. It is susceptible to the limitation of grayscale spatial distribution and is difficult to be used for MR images with variable and complex morphology. In this paper, a novel multilevel thresholding segmentation approach with a non-histogram using a modified threshold score (MTS) is proposed. An opposition-based learning hybrid rice optimization (OHRO) algorithm is used to reduce the computational cost of MTS for the purpose of optimizing the threshold search. The strategy of opposition-based learning expands the space of feasible solutions and avoids the search from stalling. The proposed approach is evaluated through the Harvard Medical School's whole brain atlas dataset. Comparing the results with TS-OHRO, Tsallis-OHRO, Kapur-OHRO, and Masi-OHRO, MTS-OHRO achieves better quantitative and qualitative outcomes which demonstrate that the application of MTS-OHRO to MR images is effective and feasible.     

Development of automatic glioma brain tumor detection system using deep convolutional neural networks

We have developed six convolutional neural network (CNN) models for finding optimal brain tumor detection system on high-grade glioma and low-grade glioma lesions from voluminous magnetic resonance imaging human brain scans. Glioma is the most common form of brain tumor. The models are constructed based on the different combinations and settings of hyperparameters with conventional CNN architecture. The six models are two layers with five epochs, five layers with dropout, five layers with stopping criteria (FLSC), FLSC and dropout (FLSCD), FLSC and batch normalization (FLSCBN), and FLSCBN and dropout. The models were trained and tested with BraTS2013 and whole brain atlas data sets. Among them, FLSCBN model yielded the best classification results for brain tumor detection. Experimental results revealed that our deep learning approach was better than the conventional state-of-art methods.     

Fully automatic method for segmentation of brain tumor from multimodal magnetic resonance images using wavelet transformation and clustering technique

Fully automatic brain tumor segmentation is one of the critical tasks in magnetic resonance imaging (MRI) images. This proposed work is aimed to develop an automatic method for brain tumor segmentation process by wavelet transformation and clustering technique. The proposed method using discrete wavelet transform (DWT) for pre‐ and post‐processing, fuzzy c‐means (FCM) for brain tissues segmentation. Initially, MRI images are preprocessed by DWT to sharpen the images and enhance the tumor region. It assists to quicken the FCM clustering technique and classified into four major classes: gray matter (GM), white matter (WM), cerebrospinal fluid (CSF), and background (BG). Then check the abnormality detection using Fuzzy symmetric measure for GM, WM, and CSF classes. Finally, DWT method is applied in segmented abnormal region of images respectively and extracts the tumor portion. The proposed method used 30 multimodal MRI training datasets from BraTS2012 database. Several quantitative measures were calculated and compared with the existing. The proposed method yielded the mean value of similarity index as 0.73 for complete tumor, 0.53 for core tumor, and 0.35 for enhancing tumor. The proposed method gives better results than the existing challenging methods over the publicly available training dataset from MICCAI multimodal brain tumor segmentation challenge and a minimum processing time for tumor segmentation. © 2016 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 26, 305–314, 2016     

Automated and reliable brain radiology with texture analysis of magnetic resonance imaging and cross datasets validation

Reliable brain tumor radiology is one of the serious mortality issues of medical hospitals and on priority of healthcare departments. In this research, the presence of brain tumor and its type (if exists) is automatically diagnosed from magnetic resonance imaging (MRI). The first step is most important where suitable parameters from Gabor texture analysis are extracted and then classified with a support vector machine. The drive of this research activity is to verify robustness of the proposed model on cross datasets, so that it could deal with variability and multiformity present in MRI data. Further to this, the developed approach is able to deploy as a real application in the local environment. Therefore, once a model has been trained and tested on an openly available benchmarked dataset, it is retested on a different dataset acquired from a local source. Standard evaluation measures, that is, accuracy, specificity, sensitivity, precision, and AUC-values have been used to evaluate the robustness of the proposed method. It has been established that the proposed method has the ability to deal with multiformity, variability, and local medical traits present in brain MRI data.     

A deep learning model integrating convolution neural network and multiple kernel K means clustering for segmenting brain tumor in magnetic resonance images

In medical imaging, segmenting brain tumor becomes a vital task, and it provides a way for early diagnosis and treatment. Manual segmentation of brain tumor in magnetic resonance (MR) images is a time‐consuming and challenging task. Hence, there is a need for a computer‐aided brain tumor segmentation approach. Using deep learning algorithms, a robust brain tumor segmentation approach is implemented by integrating convolution neural network (CNN) and multiple kernel K means clustering (MKKMC). In this proposed CNN‐MKKMC approach, classification of MR images into normal and abnormal is performed by CNN algorithm. At next, MKKMC algorithm is employed to segment the brain tumor from the abnormal brain image. The proposed CNN‐MKKMC algorithm is evaluated both visually and objectively in terms of accuracy, sensitivity, and specificity with the existing segmentation methods. The experimental results demonstrate that the proposed CNN‐MKKMC approach yields better accuracy in segmenting brain tumor with less time cost.     

Fully automatic multisegmentation approach for magnetic resonance imaging brain tumor detection using improved region-growing and quasi-Monte Carlo-expectation maximization algorithm

Magnetic resonance imaging (MRI) is widely used in the medical field, especially for detecting serious abnormalities affecting the organs of the human body, such as tumors. Automatic detection of tumors needs high-performance recognition techniques. In this paper, we have developed a new automatic method based on the multisegmentation of brain tumor region. We used an improved region-growing algorithm, which is based on quasi-Monte Carlo and expectation maximization methods to define the desired classes. Several metrics were calculated to evaluate the performance of our technique. The fully automatic multisegmentation approach, developed in this study, showed good performance, and it can offer a new option to replace conventional techniques used for tumor detection in MRI images.     

Segmentation for brain magnetic resonance images using dual‐tree complex wavelet transform and spatial constrained self‐organizing tree map

In brain MR images, the noise and low‐contrast significantly deteriorate the segmentation results. In this paper, we introduce a novel application of dual‐tree complex wavelet transform (DT‐CWT), and propose an automatic unsupervised segmentation method integrating DT‐CWT with self‐organizing map for brain MR images. First, a multidimensional feature vector is constructed based on the intensity, low‐frequency subband of DT‐CWT, and spatial position information. Then, a spatial constrained self‐organizing tree map (SCSOTM) is presented as the segmentation system. It adaptively captures the complicated spatial layout of the individual tissues, and overcomes the problem of overlapping gray‐scale intensities for different tissues. SCSOTM applies a dual‐thresholding method for automatic growing of the tree map, which uses the information from the high‐frequency subbands of DT‐CWT. The proposed method is validated by extensive experiments using both simulated and real T1‐weighted MR images, and compared with the state‐of‐the‐art algorithms. © 2014 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 24, 208–214, 2014     

Targeted nanoparticles for the non-invasive detection of traumatic brain injury by optical imaging and fluorine magnetic resonance imaging

Necrosis is a form of cell death that occurs only under pathological conditions such as ischemic diseases and traumatic brain injury (TBI). Non-invasive imaging of the affected tissue is a key component of novel therapeutic interventions and measurement of treatment responses in patients. Here, we report a bimodal approach for the detection and monitoring of TBI. PEGylated poly(lactic-co-glycolic acid) (PLGA) nanoparticles (NPs), encapsulating both near infrared (NIR) fluorophores and perfluorocarbons (PFCs), were targeted to necrotic cells. We used cyanine dyes such as IRDye 800CW, for which we have previously demonstrated specific targeting to intracellular proteins of cells that have lost membrane integrity. Here, we show specific in vivo detection of necrosis by optical imaging and fluorine magnetic resonance imaging (¹⁹F MRI) using newly designed PLGA NP(NIR700 + PFC)-PEG-800CW. Quantitative ex vivo optical imaging and ¹⁹F MR spectroscopy of NIR-PFC content in injured brain regions and in major organs were well correlated. Both modalities allowed the in vivo identification of necrotic brain lesions in a mouse model of TBI, with optical imaging being more sensitive than ¹⁹F MRI. Our results confirm increased blood pool residence time of PLGA NPs coated with a PEG layer and the successful targeting of TBI-damaged tissue. A single PLGA NP containing NIR-PFC enables both rapid qualitative optical monitoring of the TBI state and quantitative 3D information from deeper tissues on the extent of the lesion by MRI. These necrosis-targeting PLGA NPs can potentially be used for clinical diagnosis of brain injuries.

Open image in new window

     

Modified Global Flower Pollination Algorithm‐based image fusion for medical diagnosis using computed tomography and magnetic resonance imaging

Recently, the computed tomography (CT) and magnetic resonance imaging (MRI) medical image fusion have turned into a challenging issue in the medical field. The optimal fused image is a significant component to detect the disease easily. In this research, we propose an iterative optimization approach for CT and MRI image fusion. Initially, the CT and MRI image fusion is subjected to a multilabel optimization problem. The main aim is to minimize the data and smoothness cost during image fusion. To optimize the fusion parameters, the Modified Global Flower Pollination Algorithm is proposed. Here, six sets of fusion images with different experimental analysis are evaluated in terms of different evaluation metrics such as accuracy, specificity, sensitivity, SD, structural similarity index, feature similarity index, mutual information, fusion quality, and root mean square error (RMSE). While comparing to state‐of‐art methods, the proposed fusion model provides best RMSE with higher fusion performance. Experiments on a set of MRI and CT images of medical data set show that the proposed method outperforms a very competitive performance in terms of fusion quality.     

Three stage cervical cancer classifier based on hybrid ensemble learning with modified binary PSO using pretrained neural networks

ABSTRACT

Cervical cancer is one of the major challenges in developing nations like India.In recent years, a lot of research has been done todetect cervical cancer at an early stage through the pap-smear test, human papillomavirus test (HPV), etc. In this study, we have proposed athree-stage cervical cancer classifier to classify cervical cells among normal and abnormal cells using a hybrid ensemble classifier based onfeatures extracted using pre-trained neural networks. Furthermore, this work extends to classify the cells among different levels of dysplastic mainly mild, moderate and severe. The accuracy achieved for 2-class classification among normal and abnormal cells is up to 100% while for 4-class classification among normal, mild, moderate and severe dysplastic cells is up to 98.91% and 99.12% for new and old Herlev university hospital datasets respectively.