2.4. Evaluation

The performance of three video compression algorithms as mentioned above was investigated in this study. They are Motion JPEG 2000 (MJ2), Motion JPEG AVI (AVI), and MPEG‐4 (MP4). The function of three algorithms is provided by VideoWriter of Matlab (MathWorks, Inc. Natick, MA, USA). For MJ2, compression ratio is not specified in advance to allow video data to be compressed as much as possible. For AVI and MP4, the video quality is set to its maximum number 100 to allow the best quality of video. These tests were performed on a personal computer equipped with Intel i7 CPU 2.4‐GHz and 12 GB RAM. The programs for data processing were developed with Matlab (version 2013). The CBCT images were collected from 30 patients with treatment sites at head and neck (10 cases), thorax (10 cases), and pelvis (10 cases). For each patient, 8‐10 CBCT sets are used for testing.

The performance of video compression algorithm was evaluated by compression ratio and compression time. Compression ratio is defined as the ratio between the file sizes of image sequence and video file. Compression time is the average time for processing one image and calculated by the total compression time dividing the total number of images processed.

The similarity in quality of image sequence is evaluated by difference and correlation between all successive images in a sequence. Higher value of similarity of a sequence means there is more redundant information to be reduced and larger compression ratio is expected. The image difference (DIFF) is calculated by mean value of image differences in a sequence as defined below.

The image correlation (CORR) is calculated by mean value of image correlation coefficients in a sequence as defined below.

The performance of video decompression algorithm is evaluated by decompression time, mean square error (MSE), peak signal‐to‐noise ratio (PSNR), and video quality matrix (VQM). Decompression time is the average time for processing one image from video. The MSE is calculated by comparing original and decompressed images pixel by pixel as defined below.

It is important to compare the error of an image with respect to the amount of bits a pixel is encoded. PSNR is the ratio between the maximum power of a signal and the power of corrupting noise that affects the fidelity of its representation. In this case it is defined as:

here, MAX is the maximum possible pixel value of the image and 2¹⁶‐1 in this study. Typical values for the PSNR of lossy image and video compression are between 60 and 80 dB, provided the bit depth is 16 bits. VQM is a metric to predict human‐perceived video quality and is defined as:

The values of a = 0.15 and b = 19.7818 have been set experimentally. The resulting VQM is compared to fuzzy results like “excellent” (VQM < 20%) or “good” (VQM < 40%).²²

The impact of image loss on positioning accuracy was assessed using a clinical image registration application — offline review (Varian medical system, Palo Alto, CA, USA). First, the original CBCT images were automatically registered with planning CT to determine target offset for patient positioning. Next, CBCT images were compressed to a video and then decompressed from video to another set of CBCT images. The CBCT images after decompression were automatically matched with planning CT to determine another target offset. The difference between both sets of target offsets is the discrepancy caused by image loss due to compression algorithm. This discrepancy represents the inconsistency of registration accuracy before and after compression. The image registration was performed automatically and the parameters were set for bony structures and soft tissues, respectively, as shown in dialog window of Figs. 4 and ​and5.5. Specifically, the intensity ranges for bony structure and soft tissues were set to 0–200 and 200–3000. For each session, the target offsets in three dimensions were displayed at the left‐bottom corner of main application.

The illustration of cone‐beam computerized tomography registration for lung tumor patient based on bony structures.

The illustration of cone‐beam computerized tomography registration for lung tumor patient based on soft tissues.