A Data Compression Device for 4D CBCT
W Ren*, J Dai,SU-E-I-16 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
Purpose: 4D Cone beam computed tomography (CBCT) produce hundreds of images daily which are useless for diagnosis. A key problem is to effectively compress these similar imaging data daily and easily invoke them when necessary. The aim of this study is to develop a data compression device for CBCT with appropriate compressing ratio and reserving the image quality.
Methods: Kilovoltage (kV) CBCT image from one patient was acquired for each treatment fraction to correct patient setup. The transverse CBCT images were reconstructed from the raw data using a XVI system. The compression device was written in C++ using dicomtk library, allowing a selectable tradeoff between storage size and image quality, then Qt software which is a cross-platform application and UI framework was used to package this compression algorithm. In this study, JPEG (Joint Photographic Experts Group) algorithm was employed since it is the most commonly method for imaging processing.
Results: After compression, the original 1.53GB file was compressed to 461.1MB for lossless while 264.3MB for lossy compression. To evaluate the impact of the lossy compression on image quality, the compressed image and the original image were compared by using PSNR (Peak signal noise ratio), there were no obvious distortions between them.
Conclusion: A new effective data compression device for CBCT is developed without apparent influence on the image quality, and this could improve the efficiency of daily 4D CBCT and fascinate other advanced treatment techniques like adaptive tumor detection and tracking.