Intracranial hemorrhage detection in 3D computed tomography images using a bi-directional long short-term memory network-based modified genetic algorithm

• DOI: 10.3389/fnins.2023.1200630
• Publication Date: 2023-07-04T06:46:48Z
• Authors: Jewel Sengupta; Robertas Alzbutas; Przemysław Falkow...; Bożena Falkowska-...
• Keywords: intracranial hemorrhage detection; gradient local ternary pattern; bi-directional long short-term memory network; genetic algorithm; computed tomography; Tamura features; Neurosciences. Biological psychiatry. Neuropsychiatry; region of interest; RC321-571; Neuroscience

Intracranial hemorrhage detection in 3D Computed Tomography (CT) brain images has gained more attention the research community. The major issue to deal with CT is scarce and hard obtain labelled data better recognition results.To overcome aforementioned problem, a new model been implemented this manuscript. After acquiring from Radiological Society of North America (RSNA) 2019 database, region interest (RoI) was segmented by employing Otsu's thresholding method. Then, feature extraction performed utilizing Tamura features: directionality, contrast, coarseness, Gradient Local Ternary Pattern (GLTP) descriptors extract vectors RoI regions. extracted were dimensionally reduced proposing modified genetic algorithm, where infinite selection technique incorporated conventional algorithm further reduce redundancy within regularized vectors. selected optimal finally fed Bi-directional Long Short Term Memory (Bi-LSTM) network classify intracranial sub-types, such as subdural, intraparenchymal, subarachnoid, epidural, intraventricular.The experimental investigation demonstrated that Bi-LSTM based obtained 99.40% sensitivity, 99.80% accuracy, 99.48% specificity, which are higher compared existing machine learning models: Naïve Bayes, Random Forest, Support Vector Machine (SVM), Recurrent Neural Network (RNN), Short-Term (LSTM) network.