ABSTRACT
Introduction: Brain tumors arise due to abnormal growth of cells at any brain location with uneven boundaries and shapes. Usually, they proliferate rapidly, and their size increases by approximately 1.4% a day, resulting in invisible illness and psychological and behavioral changes in the human body. It is one of the leading causes of the increase in the mortality rate of adults worldwide. Therefore, early prediction of brain tumors is crucial in saving a patient's life. In addition, selecting a suitable imaging sequence also plays a significant role in treating brain tumors. Among available techniques, the magnetic resonance (MR) imaging modality is widely used due to its noninvasive nature and ability to represent the inherent details of brain tissue. Several computer-assisted diagnosis (CAD) approaches have recently been developed based on these observations. However, there is scope for improvement due to tumor characteristics and image noise variations. Hence, it is essential to establish a new paradigm. Methods: This paper attempts to develop a new medical decision-support system for detecting and differentiating brain tumors from MR images. In the implemented approach, initially, we improve the contrast and brightness using the tuned single-scale retinex (TSSR) approach. Then, we extract the infected tumor region(s) using maximum entropy-based thresholding and morphological operations. Furthermore, we obtain the relevant texture features based on the non-local binary pattern (NLBP) feature descriptor. Finally, the extracted features are subjected to a support vector machine (SVM), K-nearest neighbors (KNN), random forest (RF), and GentleBoost (GB). Results: The presented CAD model achieved 99.75% classification accuracy with 5-fold cross-validation and a 91.88% dice similarity score, which is higher than the existing models. Discussions: By analyzing the experimental outcomes, we conclude that our method can be used as a supportive clinical tool for physicians during the diagnosis of brain tumors.
ABSTRACT
OBJECTIVE: The objective of any multimodal medical image fusion algorithm is to assist a radiologist for better decision-making during the diagnosis and therapy by integrating the anatomical (magnetic resonance imaging) and functional (positron emission tomography/ single-photon emission computed tomography) information. METHODS: We proposed a new medical image fusion method based on content-based decomposition, Principal Component Analysis (PCA), and sigmoid function. We considered Empirical Wavelet Transform (EWT) for content-based decomposition purposes since it can preserve crucial medical image information such as edges and corners. PCA is used to obtain initial weights corresponding to each detail layer. RESULTS: In our experiments, we found that direct usage of PCA for detail layer fusion introduces severe artifacts into the fused image due to weight scaling issues. In order to tackle this, we considered using the sigmoid function for better weight scaling. We considered 24 pairs of MRI-PET and 24 pairs of MRI-SPECT images for fusion, and the results are measured using four significant quantitative metrics. CONCLUSION: Finally, we compared our proposed method with other state-of-the-art transformbased fusion approaches, using traditional and recent performance measures. An appreciable improvement is observed in both qualitative and quantitative results compared to other fusion methods.
