Research And Design Technologies Of Internet-Based Digital Microscope

The digital microscope is a reflection of modern virtual instrument technology's development. It is a combination of modern virtual instrument technology and traditional optical microscope. With the development of computer and internet technology, digital microscope system will be applied firstly on anatomy, histology, pathology, forensic medicine, clinical diagnosis and other related fields, for which optical microscope is a basic mean. And the digital microscope technology is developing continuously. Besides of bring benefits in the integration of resources, saving money, etc, the main advantages of the digital microscope is promoting remote information communication.The digital microscope technology will solve the problem of the current remote pathological diagnosis system, and promote the development of telemedicine. Also, virtual microscope technology will promote the reform of current medical teaching mode, and reduce the investment of pathology teaching. In addition, virtual microscope technology will change the current pathological section management mode in the hospital, reduce retrieval time, and save storage space. The digital microscope technology involves image mosaic, image fusion, image compression, image browsing, transmission and other advanced digital image processing technology, each of which may restrict the development of technology of virtual microscope. This paper focuses on research of software technology for the digital microscope. Combined with practical applications, in this paper, the disadvantage of current technologies has been improved and a complete virtual microscope platform has been implemented.For the image fusion part, in order to solve the disadvantage in the traditional image fusion method-depth of field fusion method for large images, combined with practical situation, this paper proposed a sequence color image fusion method based on dynamic zoning: first of all, dividing two adjacent images in the same scene into same size blocks dynamically. By comparing the definition of sub images in the same position of the two images, obtaining relatively clearer sub image and putting it into the corresponding position in the final fusion image. Repeating this method, we will get the final fusion image. In this method, we use a relevance based image definition evaluation function as the method to compare the adjacent images' clarity. This algorithm is fast, stable, and anti-interference. It is a better solution of the larger images' depth of field image fusion problem. In the replacing region, the use of the weighted transition solves the image joints problem, while dynamic dividing guarantees the integrity of images in the transition region.In the image mosaic technology, we optimize and improve the SIFT (Scale Invariant Feature Transform) algorithm, to avoid the disadvantage that the computing time is proportional to the image size. For the matching images, we unify some regions and normalize them to fixed size, thus greatly reducing the image matching points' searching time. We have successfully applied this improved method to the microscopic panorama image mosaic. For the feature point selection and image fusion, this paper improves the relative part, thus improves the image mosaic efficiency in higher degree. In addition, this paper uses feedback control and empirical value control in image mosaic process, thus successfully solves the problem in panoramic image mosaic process that there is no matching point in blank images and overcome the disadvantage in traditional image mosaic algorithms that the algorithm can not deal with blank images, thus makes the efficient, accurate and automatic image mosaic come true.For image compression and distribution, this paper studies and compares MrISD technology which based on wavelet-based image compression technology and Ermapper Company's ECW (Enhanced Compressed Wavelet) technology on compression ratio, compression time, cost, openness, etc. Finally, we introduce ECW (Enhanced Compressed Wavelet) technology into digital image compression and implemented a B/S based virtual microscope operation platform, using Ermapper Company's free SDK. Now the platform has been practically proven by the Shandong Medical University and WeiFang Medical University.