Measurement Method of Lesion Area Using Adaptive Seeded Region Growing

Abstract

In order to measure the lesion area of the medicine images, this study proposes a method of the lesion area measurement which is based on the adaptive seeded region growing algorithm. The method manages optimization treatment of the threshold on the common seeded region growing, so as to avoid the phenomena of the insufficient segmentation or the excessive segmentation of the regional growing as far as possible, and enhance the accuracy of the regional segmentation. On that basis, the analytic method will be used to measure the area of the lesions, which has a good performance in the measurement of irregular region. Therefore, the simulation in this study reveals that it can locate the lesions region rapidly and correctly, and calculate the area of the lesion accurately. Keywords-medical image;adaptive seeded region growing; seeded growing; irregular area; calculation of the area (key words) I. INTRODUCTION In modern medicine, it is doctors' regular work to measure and calculate the lesions area. Doctors have been using the traditional method to measure and identify the regional boundary of the heart and thoracic by using pens, rulers and films. After that they figure the approximate value by hand or with the help of calculator, and then to make the diagnosis. However, this traditional method is time-consuming and it often distains the film, and sometimes it is inaccurate. Fortunately, with the development of computer image processing technology, people find a new means for the accurate calculation of lesions area of medical images (1-3). What's more, this new method improves the work efficiency and quality of the image diagnosis. In this study, we will introduce the measurement of the lesions area based on the adaptive seeded region growing algorithm (4). By using this, the things doctors need to do is click the lesions region they want to measure, and the algorithm can find the regional boundary automatically. Besides, some modification has been made on optimization treatment which can reduce the probability of insufficient segmentation and excessive segmentation, and is better than general region growing method. What's more, it improves the accuracy of lesions region and can measure and calculate the lesions area rapidly and accurately. The biggest advantage of this method is that it greatly improves doctor's work efficiency and avoids contamination of the films at the same time. II. THE PRINCIPLE OF REGIONAL GROWING In this study, we take the method of seeded region growing (SRG). Region growing segmentation is a kind of ancient segmentation method in image segmentation, and it was raised by Adams and Bischof (5). The basic idea is that putting the pixels together to constitute a region, all pixels in region has the similar quality and characteristics. Begin with original region (center on selected seeds' pixel value), and expand region gradually by merging the neighboring pixel with similar quality and characteristics or other constituted regions to the current region. This progress continues until no pixel satisfying the condition is available.