| Summary: | 碩士 === 國立中興大學 === 資訊管理學系所 === 102 === In the global aging trend, stroke is one of the most affected diseases. It is mainly due to cerebral vascular obstruction or bleeding, making the brain blood circulation disorder and lack of nutrients and oxygen, resulting in brain damage or death then producing a variety of neurological symptoms. This study uses the rat model of middle cerebral artery occlusion as experimental subjects, and takes NeuN antibody staining rat brain neurons smear images for the study images. The stroke diagnostic system analyzes the texture of the images by gray level co-occurrence matrix and Tamura, further combines F-values to select the textural features used in final, and then finds the phase feature of each phase, finally decides the optimal feature weights by genetic algorithm based parameter detector. The experimental results show that the can accurately diagnose the stroke phase of NeuN antibody staining rat brain neurons smear images, and also useful to evaluate therapeutic effect of treatments.
|
|---|
