A Study On The Application Of Compression And 3-D Reconstruction For Images Of Multifocal Multiphoton Microscopy

Multifocal multiphoton microscopy is one of the most powerful microscopic technologies for in vivo 3D tissue imaging. Its near-infrared light is more penetrating, own minimal photodamage and photobleaching and become important applications in neuronal imaging, tumor physiology study, and so on. Three-dimensional reconstruction and image compression of MMM imaging are important in its application field. By reconstructing the slices of MMM, the slices generate three-dimensional images which researchers needed in real-time, so that researchers can analyze and study biological tissues. Its better resolution leads that the image data is huge amount. Image compression can be used to compress the huge amount of data for image data storage and transmission.In terms of the characteristics and limitations of MMM, the collected biological tissue slice data have noise or other bad effects inevitably. The study of image preprocessing can improve the image quality. Image preprocessing of MMM involve image filtering and gray correction. The filter effects were made the comparison by using mean filter, median filter and low pass filter. A kind of gray correction approach was proposed based on region growing to segment the image background block from the target tissues and get Region of Interest. Three-dimensional reconstruction techniques include two classes - volume rendering and surface rendering. These theories and methods were described in detail that are optical models, basic principles and implementation processes of the Ray Casting algorithm and Marching Cubes algorithm, and made a comparison of rendering effects by using the sample data. These methods are ready for the future system development. In aspect of image compression, lossy compression and lossless compression were used, such as DCT-based JPEG and Huffman coding. In experiments of image compression, the compressed result of two methods were analyzed and made the comparison and evaluation.Based on previous theoretical study, MMM image compression and 3D reconstruction system was developed by using the three-dimensional reconstruction algorithm of Medical Image ToolKit of Institute of Automation Chinese Academy of Sciences and using MFC to build the application framework. Finally, the sample data was used to do some experiments in the system.In comparison with MC algorithm, volume rendering based on RC algorithm is more suitable for sample data of multifocal multiphoton microscopy. In aspect of image compression, Huffman coding is better than the performance of JPEG.