فشرده سازي بدون اتلاف نواحي با اهميت تصاوير آنژيوگرافي قلب

چكيده انگليسي :

Lossless Compression of ROI in Coronary Angiogram Sequences Zahra Saffar Yazdi z saffaryazdi@ec iut ac ir March 8 2010 Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan 84156 83111 Iran Degree M Sc Language FarsiS Samavi Prof Supervisor samavi96@cc iut ac irM Davarpanah Jazi Assist Prof Supervisor mdjazi@cc iut ac irAbstract One of the most important reasons of heart attack is the narrowing of arteries Coronary angiography is anefficient x ray examination for diagnosing coronary artery diseases and diagnosing blocked arteries Also theabnormalities in the blood vessels such as stenosis can be diagnosed During the angiography procedure anarrow tube that is called a catheter enters the vessel from groin region After placing catheter in the heart orin the entrance of the feeding arteries of the heart a contrast agent is injected through Then a number of X Rayimages which are called angiogram are taken Blood vessels become visible after the injection and an expertcan diagnose coronary artery diseases without the need for a hard invasive operation For a patient this procedure takes 4 to 5 minutes during which about 4000 images are taken The size of theseimages is typically 512 512 pixels Usually these images are taken at a speed of 30 frames per second with atleast 8 bit pixels The size of the raw data in this procedure is approximately 2 5 GB Storage and transmissionof this large amount of data is a challenge for hospitals Efficient compression of these data files is an inevitablerequirement Image sequences from digital angiography contain areas of high diagnostic interest Loss of information dueto compression for regions of interest ROI in angiograms is not tolerable Due to the sensitivity of the medicaldata in medical images lossless methods are preferred But the maximum compression ratio in these methods isabout 2 or 3 1 The compression ratio in the Lossy methods is always higher than those of the lossless methods On the other hand in the lossy methods some important medical details may be lost and diagnostic errors mayoccur By analysing angiogram images it is concluded that large areas of each frame does not containdiagnostically important information Therefore it is concluded that using of lossy methods for some parts of animage can be medically acceptable In the literature a lot of work is done on the lossless and lossy compression of the angiogram images Someof these researches identify a region of interest ROI and allocate more bit budget for the diagnosticallyimportant regions These methods are based on both 2D and 3D techniques similar to the video compressionmethods Researchers compared inter frame and intra frame methods such as wavelet DCT and motion estimationalgorithm in a study Then they presented a lossy method that first removes spatial correlation by using wavelet and then temporal correlation by motion estimation Also researchers have presented an ROI based methodwhere by modelling image motions the image is segmented into diagnostically important and diagnosticallynon important regions Then the images are compressed with a 3D SPIHT algorithm In mentioned method theROI is compressed with more accuracy In another study researchers have replaced DCT with wavelets inMPEG and compressed these images in a lossy manner Also a modified H 264 video compression methodapplied for the compression of the 3D Angiogram images In another research a lossless method that workedbased on the motion estimation and a context adaptive predictor was presented In this thesis both lossy and lossless techniques are applied to the angiogram images In these images thereare regions where no vessels are present Therefore preserving all of the details of these regions would not servethe diagnostic procedures In our method we either do not store any data for some of these regions or encodethem in a lossy manner For regions where arteries are present lossless encoding is applied In our method first of

صفاريزدي، زهرا

فشرده سازي بدون اتلاف نواحي با اهميت تصاوير آنژيوگرافي قلب

تطابق بلاك , مدل سازي متن