Abstract-Textural feature based classification has shown that magnetic resonance images can characterize histological brain tumor types. Feature selection is an important process to acquire a robust textural feature subset and enhance classification rate. This work investigates two different feature selection techniques; principal component analysis (PCA), and the combination of max-relevance and min-redundancy (mRMR) and feedforward selection. We validated these techniques based on a multi-center dataset of pediatric brain tumor types; medulloblastoma, pilocytic astrocytoma and ependymoma, and investigated the accuracy of tumor classification, based on textural features of diffusion and conventional MR images.
A specific design of craniofacial implant model is vital and urgent for patients with traumatic head injury. The mirror technique is commonly used for modeling these implants, but it requires the presence of a healthy skull region opposite to the defect. To address this limitation, we propose three processing workflows for modeling craniofacial implants: the mirror method, the baffle planner, and the baffle-based mirror guideline. These workflows are based on extension modules on the 3D Slicer platform and were developed to simplify the modeling process for a variety of craniofacial scenarios. To evaluate the effectiveness of these proposed workflows, we investigated craniofacial CT datasets collected from four accidental cases. The designed implant models were created using the three proposed workflows and compared to reference models created by an experienced neurosurgeon. The spatial properties of the models were evaluated using performance metrics. Our results show that the mirror method is suitable for cases where a healthy skull region can be completely reflected to the defect region. The baffle planner module offers a flexible prototype model that can be fit independently to any defect location, but it requires customized refinement of contour and thickness to fill the missing region seamlessly and relies on the user's experience and expertise. The proposed baffle-based mirror guideline method strengthens the baffle planner method by tracing the mirrored surface. Overall, our study suggests that the three proposed workflows for craniofacial implant modeling simplify the process and can be practically applied to a variety of craniofacial scenarios. These findings have the potential to improve the care of patients with traumatic head injuries and could be used by neurosurgeons and other medical professionals.
Quantification of parasitaemia is an important part of a microscopic malaria diagnosis. Giemsastained thin blood smear is the gold standard method for detecting malaria parasite enumeration. However, manual counting reveals the limitations of human inconsistency and fatigue, as well as the unreliability of accuracy and non-reproducibility. In this paper, the texture-based classification approach is investigated. The methods consist of the following processes: pre-processing, segmentation, feature extraction and the classification of erythrocytes. The preprocessing is applied for image conversion and enhancement. The segmentation combines local adaptive thresholding, morphological process and watershed transform to extract red blood cells, separate touching and overlapping cells. Texture analysis is performed to establish parameters obtained from first-order, second-order and higher-order statistical analysis and wavelet transform. Two feature selection approaches, the sequential forward selection method and sequential backward selection method, integrated with a support vector machine classifier are examined to obtain the optimal feature set for identifying the Plasmodium falciparum stages. We found that graylevel co-occurrence matrices based textural features were highly selected. The proposed method produces 98.87% accuracy for binary classification, 99.56% accuracy for ring stage classification, and 99.48% accuracy for tropozoite stage classification.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.