Ct Images Of The Heart Classification Retrieval Techniques

CT Heart Image Retrieval is an important part of CT Image Retrieval. How to retrieve a heart image from a heart image database which contains a large number of CT images is a key problem of the assistant diagnose of heart disease using medical image. Medical CT image is mainly expressed in gray-scale information. So, extracting gray-scale histogram is particularly important for medical image retrieval. But the gray-scale information of CT heart image is not evident compared to other images. The traditional medical retrieval systems only use gray-scale histogram for searching CT image and the results are not accurate. Therefore this article researched and designed the technology of CT heart image retrieval using gray-scale histogram and shape characteristics.As the gray-scale histogram has some characteristics such as ratio, rotation and invariance of displacement, we can extract gray-scale information by using gray-scale histogram. With the combination of gray-scale transformation and gray-scale histogram, we realized the primary search of gray-scale histogram. However, gray-scale histogram can not fully express the spatial distribution information. The shape characteristics can overcome this shortcoming. In this paper, the edge and the region of the CT heart images was described. We used centroid distance increment matrix to extract the CT heart shape characteristics. With the result of edge extraction and segmentation, we extracted the characteristics of the border image in each scale. These characteristics composed the feature vector matrix of the image. So the heart shape diagram was built and the similarity was evaluated by using euclidean distance. Then the CT heart image was searched eventually. The experiments showed that this method is stable, general and lower error, and it can search the CT heart image rightly.