2024-03-29T06:58:46Zhttps://eprints.lib.hokudai.ac.jp/dspace-oai/requestoai:eprints.lib.hokudai.ac.jp:2115/398282022-11-17T02:08:08Zhdl_2115_39595hdl_2115_39594hdl_2115_33096Medical Images Illustrator : A Flexible Image Processing SystemKu, Cheng-WeiTsai, Kuen-LongLi, Po-Ying548In this paper, we propose a medical image processing and visualization system, Medical Image Illustrator (MIIL), which is able to help doctors for symptom diagnosis, detection and surgical planning. MIIL can process and visualize the Digital Imaging and Communications in Medicine (DICOM) data directly. We have implemented some processing concepts, such as "dynamic processing flow", "dynamic flowing branch" and "multiple masks / images operation", which make MIIL a very flexible workbench for image processing. The system can also be easy extendable for different situations. Dynamic processing flow enables MIIL to apply a variety of 2D and 3D image processing algorithms on images or masks to filter out some features. This can results in better images for analysis and segmentation. Dynamic flowing branch allows the user to create different process-flow on images or masks. Each flowing branch can apply different operations independently. Users can also perform operations on multiple flows, or combine the results from different processing-flow. In the visualization stage, the 3D volume datasets are generated from different masks of 3D images. MIIL supports interactive direct volume rendering on multiple datasets, users can visualize these data with the ability to adjust the transfer function of each dataset individually. Therefore, MIIL could allow users to visualize and interact with multiple volumes that each volume represents different tissue or organ. Besides, MIIL also supports stereoscopic display, which is suitable for Virtual Reality (VR) application to enhance the visual effect. With above features, MIIL can provide very useful and decent visual effect to display medical or other 3D images. We will give two examples which use MIIL to process data. In the first case, we use MIIL to process the CT data of human chest. We were able to find possible tumor tissues in the data and provide visual analytics to these tissues. In the second case, we use MIIL to process the T1 MRI and Diffusion Spectrum MRI (DSI) together, doctors are able to visualize the relationship between tumor, tracts and brain structure. Based on the above two successful cases, we believe that MIIL is helpful for diagnosis and surgical planning for a trained medical staff.Asia-Pacific Signal and Information Processing Association, 2009 Annual Summit and Conference, International Organizing CommitteeConference Paperapplication/pdfhttp://hdl.handle.net/2115/39828https://eprints.lib.hokudai.ac.jp/dspace/bitstream/2115/39828/1/WA-L4-1.pdfProceedings : APSIPA ASC 2009 : Asia-Pacific Signal and Information Processing Association, 2009 Annual Summit and Conference8878942009-10-04engpublisher