Efficiently compressing 3D medical images for teleinterventions via CNNs and anisotropic diffusion

Medical Physics - Tập 48 Số 6 - Trang 2877-2890 - 2021
Luu Manh Ha1,2,3, Theo van Walsum2, Daniel Franklin4, Phuong Cam Pham5, Luu Dang Vu6, Adriaan Moelker2, Marius Staring7, Xiem VanHoang3, Wiro J. Niessen2, Nguyen Linh-Trung1
1AVITECH, University of Engineering and Technology, VNU, Hanoi, Vietnam
2Department of Radiology and Nuclear Medicine, Erasmus MC, Rotterdam, The Netherlands
3FET, University of Engineering and Technology, VNU, Hanoi, Vietnam
4School of Electrical and Data Engineering, University of Technology Sydney, Sydney, Australia
5Nuclear Medicine and Oncology Center, Bach Mai Hospital, Hanoi, Vietnam
6Radiology Center, Bach Mai Hospital, Hanoi, Vietnam
7Department of Radiology, Leiden University Medical Center, Leiden, The Netherlands

Tóm tắt

Purpose

Efficient compression of images while preserving image quality has the potential to be a major enabler of effective remote clinical diagnosis and treatment, since poor Internet connection conditions are often the primary constraint in such services. This paper presents a framework for organ‐specific image compression for teleinterventions based on a deep learning approach and anisotropic diffusion filter.

Methods

The proposed method, deep learning and anisotropic diffusion (DLAD), uses a convolutional neural network architecture to extract a probability map for the organ of interest; this probability map guides an anisotropic diffusion filter that smooths the image except at the location of the organ of interest. Subsequently, a compression method, such as BZ2 and HEVC‐visually lossless, is applied to compress the image. We demonstrate the proposed method on three‐dimensional (3D) CT images acquired for radio frequency ablation (RFA) of liver lesions. We quantitatively evaluate the proposed method on 151 CT images using peak‐signal‐to‐noise ratio (), structural similarity (), and compression ratio () metrics. Finally, we compare the assessments of two radiologists on the liver lesion detection and the liver lesion center annotation using 33 sets of the original images and the compressed images.

Results

The results show that the method can significantly improve of most well‐known compression methods. DLAD combined with HEVC‐visually lossless achieves the highest average of 6.45, which is 36% higher than that of the original HEVC and outperforms other state‐of‐the‐art lossless medical image compression methods. The means of and are 70 dB and 0.95, respectively. In addition, the compression effects do not statistically significantly affect the assessments of the radiologists on the liver lesion detection and the lesion center annotation.

Conclusions

We thus conclude that the method has a high potential to be applied in teleintervention applications.

Từ khóa


Tài liệu tham khảo

SrivastavaS PantM AbrahamA AgrawalN.The Technological Growth in eHealth Services 2015 ISSN: 1748‐670X Pages: e894171 Publisher: Hindawi Volume: 2015.

10.1016/j.jacr.2019.05.053

10.1007/s13244-012-0210-z

10.1007/s13244-016-0485-6

10.3389/fpubh.2014.00125

Hanoi Medical University Hospital ‐ HMU Hospital ‐ YouTube.https://www.youtube.com/watch?v=Nyu9D1y‐HnI. Accessed September 24 2020.

SingLIVE 2020: Scaling Greater Heights ‐ SingHealth.https://www.sgh.com.sg/news/education/singlive‐2020‐scaling‐greater‐heights;https://www.emedevents.com/c/medical‐conferences‐2020/singapore‐live‐2020;https://www.nhcs.com.sg/about‐us/newsroom/Documents/NHCS_Murmurs_Issue36‐WEB.pdf. Accessed September 24 2020.

KaufmanJA LeeMJ Vascular and Interventional Radiology. The Requisites E‐Book Elsevier Health Sciences.2013.https://www.elsevier.com/books/vascular‐and‐interventional‐radiology‐the‐requisites/kaufman/978‐0‐323‐04584‐1

10.3389/fpubh.2019.00244

10.1016/j.ejrad.2009.06.025

AhadA TahirM YauK‐LA.5G‐Based Smart Healthcare Network: Architecture Taxonomy Challenges and Future Research Directions IEEE Access 7 Conference Name: IEEE Access;2019:100747–100762.

LammleT.CompTIA Network+ Study Guide: Exam N10–007 4th edn. Comptia Network + Study Guide Authorized Courseware Sybex;2018.

KumarS KrupinskiE Teleradiology Springer Science & Business Media 2008.

ClunieDA.Lossless compression of grayscale medical images: effectiveness of traditional and state‐of‐the‐art approaches in Medical Imaging 2000: PACS Design and Evaluation: Engineering and Clinical Issues volume 3980 International Society for Optics and Photonics.2000:74–84.

10.1016/S0262-8856(00)00064-0

ParikhS KalvaH AdzicV.Evaluation of HEVC compression for high bit depth medical images in 2016 IEEE International Conference on Consumer Electronics (ICCE) 2016:311–314 ISSN: 2158‐4001.

UmaMaheswari S, 2020, Lossless medical image compression algorithm using tetrolet transformation, J Ambient Int Human Comput, 1

10.1007/s11265-016-1150-5

10.1016/j.image.2017.02.002

10.1109/TMI.2017.2714640

HulskenB.Fast Compression Method for Medical Images on the Web arXiv:2005.08713 [eess] (2020) arXiv: 2005.08713.https://arxiv.org/abs/2005.08713

10.5858/2000-124-1653-EOICOT

10.1109/JBHI.2017.2660482

10.1007/s10278-019-00283-3

10.2352/ISSN.2470-1173.2020.10.IPAS-063

10.1186/s40064-016-3784-y

10.1504/IJCVR.2016.073759

10.1007/978-981-15-2043-3_32

10.1109/TMI.2009.2013851

10.1016/j.media.2017.07.005

AhmadiM EmamiA HajabdollahiM et al.Lossless Compression of Angiogram Foreground with Visual Quality Preservation of. Background. 2018;arXiv:1802.07769. arXiv:1802.07769 [cs].

10.1148/radiol.2281020254

MoorthiM AmuthaR.A near Lossless compression method for medical images in IEEE‐International Conference On Advances In Engineering. Science And Management (ICAESM ‐2012) 2012:39–44.

10.1007/s12046-013-0126-4

10.1080/03772063.2017.1309998

10.1016/j.procs.2015.10.037

10.1007/s10916-018-1090-7

10.1002/j.1538-7305.1948.tb01338.x

10.1055/s-0035-1564705

10.1097/MD.0000000000001594

LuuHM.Image Analysis for Guidance in Minimally Invasive Liver Interventions Number: 369.2017.

ChristPF EttlingerF GrünF et al.Automatic Liver and Tumor Segmentation of CT and MRI Volumes using Cascaded Fully Convolutional. Neural Networks.2017;arXiv:1702.05970. arXiv:1702.05970 [cs].

Son HH, 2020, Liver segmentation on a variety of computed tomography (CT) images based on convolutional neural networks combined with connected components, VNU J Sci, 36, 1

10.1007/978-94-017-1699-4_3

TariqZB ArshadN NabeelM.Enhanced LZMA and BZIP2 for improved energy data compression in 2015 International Conference on Smart Cities and Green ICT Systems (SMARTGREENS);2015:1–8.

PrangnellL Visually lossless coding in HEVC: A high bit depth and 4: 4: 4 capable JND‐based perceptual quantisation technique for HEVC Signal Processing: Image Communication2018;63:125–140.

RonnebergerO FischerP BroxT U‐Net: Convolutional Networks for Biomedical Image Segmentation 2015; arXiv:1505.04597. arXiv:1505.04597 [cs].

HoangHS Phuong PhamC FranklinD vanWalsumT Ha LuuM.An Evaluation of CNN‐based Liver Segmentation Methods using Multi‐types of CT Abdominal Images from Multiple Medical Centers in 2019 19th International Symposium on Communications and Information Technologies (ISCIT) 2019:20–25 ISSN: 2643‐6175.

Cover TM, 1999, Elements of Information Theory

10.1007/s10278-007-9044-5

10.1109/76.499834

BurrowsM WheelerDJ.A block‐sorting lossless data compression algorithm. Technical report.1994.

10.1109/JRPROC.1952.273898

10.1145/1082036.1082039

PatelRA ZhangY MakJ DavidsonA OwensJD Parallel lossless data compression on the GPU in 2012 Innovative Parallel Computing (InPar) 2012:1–9.

10.1109/TCSVT.2012.2221191

SanchezV Bartrina‐RapestaJ.Lossless compression of medical images based on HEVC intra coding in 2014 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP);2014:6622–6626 ISSN: 2379‐190X.

PoleA ShriramR.3‐D Medical Image Compression by Using HEVC. In 2018 Fourth International Conference on Computing Communication Control and Automation (ICCUBEA);2018:1–5.

10.1109/TCSVT.2015.2406199

Luu HM, 2016, Non‐rigid registration of liver CT images for CT‐guided ablation of liver tumors, PLOS ONE 11, e0161600

10.1109/TIP.2003.819861

10.1117/1.JMI.4.3.035501

AnderssonP NilssonJ Akenine‐MöllerT OskarssonM ÅströmK FairchildMD FLIP: A Difference Evaluator for Alternating Images Proceedings of the ACM on Computer Graphics and Interactive Techniques2020;3:1–23.

10.1002/mp.12345

BilicP The Liver Tumor Segmentation Benchmark (LiTS) 2019; arXiv:1901.04056 [cs] arXiv: 1901.04056 version: 1.

10.1007/978-3-642-39360-0_3

10.1148/radiol.2018180125

BrossB ChenJ LiuS.Versatile video coding (Draft 5). JVET‐K10012018.

10.2214/AJR.10.5122

10.4061/2011/104685

Kalaiselvi T, 2018, Performance of medical image processing algorithms implemented in CUDA running on GPU based machine, Int J Intell Syst App, 10, 58

10.1109/TMI.2018.2806309

10.1002/mp.14196

WHO CORONAVIRUS DISEASE (COVID‐19) Dashboard.https://covid19.who.int/. Accessed October 19 2020.