Research On FCM Fuzzy Clustering Algorithm Based On Intracranial Hemorrhage CT Image

Cerebral hemorrhage is one of the most common symptoms of acute neurological injury in adults and is of concern because of its high mortality and poor postoperative recovery.The quantification of cerebral hemorrhage is very important for predicting the bleeding volume of bleeding lesions and planning follow-up treatment.Although the manual method is the most accurate segmentation result,this method is time-consuming and labor-intensive and may introduce variability when the data set is large.Like the case of clinical trials,a large number of datasets must be analyzed by different locations or different people,so manual and semi-automatic segmentation methods consume a lot of valuable time,and due to different methods of analysis of different people,the possibility of introducing random errors is also Higher.Therefore,in a clinical environment with limited time and resources,an automatic,accurate,and rapid segmentation method is needed.For decades,MRI has been proven to be the gold standard for quantifying cerebral hemorrhage,but MRI imaging is long and costly and therefore has lower usability than CT.Today,CT technology is one of the popular techniques for the diagnosis of cerebral hemorrhage.Because of its non-invasive,painless,rapid characteristics and high contrast between tissue and blood in CT images,CT technology provides technical support in most hospitals and emergency services.Fresh blood or bleeding lesions are brighter in the CT image than in other tissues.The sensitivity of detecting lesions within 24 hours prior to the onset of illness was 90%,which decreased to 80% in the first three days and decreased to 50% in one week.In order to quantify the volume of a hemorrhagic lesion,the lesion area must be segmented in a CT scan image.Without the anatomical model,all possible variations of different structures(eg,shape,size,texture,etc.),low signal-to-noise ratios,inherent artifacts,false shadows in the scan,and noise make the segmentation lesion challenging.The purpose of this study is to develop an accurate and automatic segmentation method for intracerebral hemorrhage lesions in clinical trials,to solve the FCM algorithm's sensitivity to noise,and to explore the suitability of FCM fuzzy clustering algorithm and its improved algorithm in the segmentation of hemorrhagic lesions.And applicability.First,the standard and structure of DICOM files are introduced in this article.An algorithm for converting DICOM images into other common image formats.The algorithm can not only display images in a common format but also save patient details and other related information.Second,preprocessing brain CT images.In the preprocessing operation,this paper proposes an improved Gaussian Laplacian operator for edge extraction of cranial CT images.The algorithm adds an edge intensity detector to the original Gaussian Laplacian operator.Pseudo edge points extract true edge points.Again,the image is subjected to an intracranial extraction operation based on a left-right scan algorithm.Finally,the bleeding lesions were segmented.In order to achieve segmentation of intracranial hemorrhage lesions,this paper proposes an improved FCM clustering algorithm.The spatial information is considered in both the definition of membership function and the definition of objective function.The experimental results show that the method can accurately and efficiently segment bleeding lesions without prior manual intervention and insensitive to noise.