基于可伸缩视频编码失真估计模型的超声图像码流抽取新方法

A New Bitstream Extraction Method for Ultrasound Images Based on Distortion Estimation Model of Scalable Video Coding

目的研究载人航天任务条件下可伸缩视频框架在超声图像编码中的应用,提供在轨超声检测图像实时传输的码流抽取方法。方法使用一种新的码流抽取算法,通过编码端的重建信息来估计漂移失真;考虑帧到帧的漂移失真传播系数,重建编码的失真估计模型。结果本文所使用的方案相比可伸缩视频编码参考模型(joint scalable video model,JSVM)使用的质量层分配方案,峰值信噪比(peak signal-to-noise ratio,PSNR)相当或略有提升,而计算量上避免了反复解码的复杂操作,对相同测试序列的平均抽取时间提高了10倍。结论本文提出的码流抽取方案,不需要复杂的解码操作,降低了计算复杂度,在载人航天超声图像等可伸缩视频应用中有一定应用前景。

Objective To research the application of scalable video coding in ultrasound image coding under the conditions of manned space mission, providing bitstream extraction method of real-time ultrasonic detection im- age transmission. Methods The reconstruction information of encoding side by a new way was used to estimate the drift distortion. Distortion model was reconstructed to estimate the distortion of reconstructed pictures with drift error while the frame to frame drift distortion propagation coefficients were considered. Results Compared with the joint scalable video model （JSVM） quality layers allocation algorithm, the peak signal-to-noise ratio （PSNR） was equal to or slightly improved, but the average extracting time of the same test sequences reduced by 10 times due to no repeated decoding operations. Conclusion The proposed bitstream extraction algorithm does not require complex decoding operation, and greatly reduces the computational complexity of encoding side. There are certain application prospects in the ultrasound images scalable video applications in manned space mission.