Background Current medical image compression techniques primarily optimize for mean squared error (MSE), which does not fully capture human subjective perception of image quality and often fail to preserve the structural features essential for clinical diagnosis. Objective To propose a low-loss compression coding algorithm for subtle features in medical images, aiming to reduce transmission bandwidth without compromising subjective image quality. Methods CT image sequences from 14 orthopedic surgeries at Chinese PLA General Hospital were collected in this study. Firstly, the Structural Similarity Index (SSIM) was reconstructed based on key visual features of medical images, including brightness, contrast, and detail texture, with the brightness factor set to α= 1.15 and the contrast/structure factors set to β=γ= 0.95. Subsequently, a relationship between the SSIM and MSE was established based on the linear distortion model and the law of large numbers. Then, 1/SSIM was employed as a distortion metric, and an SSIM-based distortion measure suitable for rate-distortion optimization (RDO) was constructed. On this basis, a SSIM-based rate-distortion optimization framework was developed by minimizing the distortion metric under a given target bitrate constraint. Finally, the proposed method was implemented on the x264 platform, and its rate-distortion performance was compared with that of the standard encoder to verify its advantages. Results Compared to the standard x264 encoder, our approach achieved an average rate-distortion gain of -5.2% under constant quantization parameter and -4.8% under constant quality factor. In terms of subjective quality, the SSIM of the encoded images remained above 0.95, with an average bitrate reduction of 372 kbps. Furthermore, no increase in computational complexity or encoding time was observed. Conclusion The proposed method effectively preserves the high perceptual quality of medical images while maintaining computational efficiency, offering a superior compression solution for medical image transmission.
随着视频编码技术的不断进步,最新编码标准如高效视频编码(high efficiency video coding,HEVC)、下一代标准通用视频编码(versatile video coding,VVC)和AV1已在医学图像、视频压缩领域展现出更优的压缩效率和视觉质量[20-22]。这些研究表明,VVC在PSNR增益和比特率节省方面都优于HEVC和AV1,对于更高的视频分辨率,AV1优于HEVC。而对于超声视频数据集使用的有效分辨率,HEVC则要优于AV1编码器。尽管本研究由于实验资源和环境限制,未能直接与这三种先进编码器进行实验对比,但所提方法依托传统编码体系并融入基于SSIM的感知优化策略,兼顾了算法复杂度和编码效率,在资源受限的实际应用环境下具备较强的可行性与推广潜力。未来工作计划引入HEVC、VVC和AV1编码框架。
Ali AbumallohR, NilashiM, Yousoof IsmailM,et al . Medical image processing and COVID-19: A literature review and bibliometric analysis[J]. J Infect Public Health,2022,15(1):75-93.
[2]
TayYX, KothanS, KadaS,et al . Challenges and optimization strategies in medical imaging service delivery during COVID-19[J]. World J Radiol,2021,13(5):102-121.
[3]
AlenoghenaCO, OhizeHO, AdejoAO,et al . Telemedicine: a survey of telecommunication technologies,developments,and challenges[J]. J Sens Actuator Networks,,2023,12(2):20.
[4]
WangYL, XiongHL, SunKC,et al . Toward general text-guided multimodal brain MRI synthesis for diagnosis and medical image analysis[J]. Cell Rep Med,2025,6(6):102182.
[5]
ChikkerurS, SundaramV, ReissleinM,et al . Objective video quality assessment methods: a classification,review,and performance comparison[J]. IEEE Trans Broadcast,2011,57(2):165-182.
[6]
LuuHM, van WalsumT, FranklinD,et al . Efficiently compressing 3D medical images for teleinterventions via CNNs and anisotropic diffusion[J]. Med Phys,2021,48(6):2877-2890.
[7]
UsmanMA, MartiniMG . On the suitability of VMAF for quality assessment of medical videos: Medical ultrasound & wireless capsule endoscopy[J]. Comput Biol Med,2019,113:103383.
[8]
TanHL, LiZG, TanYH,et al . A perceptually relevant MSE-based image quality metric[J]. IEEE Trans Image Process,2013,22(11):4447-4459.
VenugopalG, MullerK, PfaffJ,et al . Region-based template matching prediction for intra coding[J]. IEEE Trans Image Process,2023,32:779-790.
[11]
WangZ, BovikAC, SheikhHR,et al. Image quality assessment: from error visibility to structural similarity[J]. IEEE Trans Image Process,2004,13(4):600-612.
[12]
ZhangXX, SisniegaA, ZbijewskiWB,et al . Combining physics-based models with deep learning image synthesis and uncertainty in intraoperative cone-beam CT of the brain[J]. Med Phys,2023,50(5):2607-2624.
[13]
HarshalathaY, BiswasPK . Rate distortion optimization using SSIM for 3D video coding[C]//2016 23rd International Conference on Pattern Recognition (ICPR). December 4-8,2016,Cancun,Mexico. IEEE,2017:1261-1266.
[14]
ShoM . Properties of the SSIM metric in medical image assessment: correspondence between measurements and the spatial frequency spectrum[J]. Phys Eng Sci Med,2023,46(3):1131-1141.
[15]
PareekPK, SridharC, KalidossR,et al . IntOPMICM: intelligent medical image size reduction model[J]. J Healthc Eng,2022,2022:5171016.
[16]
PourasadY, CavallaroF . A novel image processing approach to enhancement and compression of X-ray images[J]. Int J Environ Res Public Health,2021,18(13):6724.
[17]
PanayidesAS, AminiA, FilipovicND,et al . AI in medical imaging informatics: current challenges and future directions[J]. IEEE J Biomed Health Inform,2020,24(7):1837-1857.
[18]
KesnerA, LaforestR, OtazoR,et al . Medical imaging data in the digital innovation age[J]. Med Phys,2018,45(4):e40-e52.
[19]
LiuF, Hernandez-CabroneroM, SanchezV,et al . The current role of image compression standards in medical imaging[J]. Information,2017,8(4):131.
[20]
PanayidesAS, PattichisMS, PantziarisM,et al . The battle of the video codecs in the healthcare domain - a comparative performance evaluation study leveraging VVC and AV1[J]. IEEE Access,2020,8:11469-11481.
[21]
WangY, TohidypourHR, PourazadMT,et al . Comparison of modern compression standards on medical images for telehealth applications[C]//2023 IEEE International Conference on Consumer Electronics (ICCE). January 6-8,2023,Las Vegas,NV,USA. IEEE,2023:1-6.
[22]
BuiV, ChangLC, LiDL,et al . Comparison of lossless video and image compression codecs for medical computed tomography datasets[C]//2016 IEEE International Conference on Big Data (Big Data). December 5-8,2016,Washington,DC,USA. IEEE,2017:3960-3962.
[23]
PambrunJF, NoumeirR . Limitations of the SSIM quality metric in the context of diagnostic imaging[C]//2015 IEEE International Conference on Image Processing (ICIP). September 27-30,2015,Quebec City,QC,Canada. IEEE,2015:2960-2963.
[24]
MudengV, KimM, ChoeSW . Prospects of structural similarity index for medical image analysis[J]. Appl Sci,2022,12(8):3754.