Evaluation of Deep Learning Methods for Bone Suppression from Dual Energy Chest Radiography

Sirazitdinov, Ilyas; Kubrak, Konstantin; Kiselev, Semen; Tolkachev, Alexey; Kholiavchenko, Maksym; Ibragimov, Bulat

doi:10.1007/978-3-030-61609-0_20

Cited by 3 publications

(3 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…DE-like imaging by our framework may help diagnose diseases (nodule, bone fracture, etc. ), as shown in Figures 13,15,and 16. For more accurate diagnostic, DE-like images, including lateral and oblique views, can provide addition information.…”

Section: Discussionmentioning

confidence: 99%

“…Numerous studies have focused on synthetic DE imaging using artificial intelligence. A few models use U-Net and multilevel wavelet convolutional neural network (CNN) [16][17][18][19] for shallow structures, which is associated with reduced processing capacity and low estimation accuracy. Considering that our methodology uses chest X-ray images acquired from various angles from CT images, these problems can increase when learning from irregular data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Framework for dual‐energy‐like chest radiography image synthesis from single‐energy computed tomography based on cycle‐consistent generative adversarial network

Lee

et al. 2023

Medical Physics

View full text Add to dashboard Cite

BackgroundDual‐energy (DE) chest radiography (CXR) enables the selective imaging of two relevant materials, namely, soft tissue and bone structures, to better characterize various chest pathologies (i.e., lung nodule, bony lesions, etc.) and potentially improve CXR‐based diagnosis. Recently, deep‐learning‐based image synthesis techniques have attracted considerable attention as alternatives to existing DE methods (i.e., dual‐exposure‐based and sandwich‐detector‐based methods) because software‐based bone‐only and bone‐suppression images in CXR could be useful.PurposeThe objective of this study was to develop a new framework for DE‐like CXR image synthesis from single‐energy computed tomography (CT) based on a cycle‐consistent generative adversarial network.MethodsThe core techniques of the proposed framework are divided into three categories: (1) data configuration from the generation of pseudo CXR from single energy CT, (2) learning of the developed network architecture using pseudo CXR and pseudo‐DE imaging using a single‐energy CT, and (3) inference of the trained network on real single‐energy CXR. We performed a visual inspection and comparative evaluation using various metrics and introduced a figure of image quality (FIQ) to consider the effects of our framework on the spatial resolution and noise in terms of a single index through various test cases.ResultsOur results indicate that the proposed framework is effective and exhibits potential synthetic imaging ability for two relevant materials: soft tissue and bone structures. Its effectiveness was validated, and its ability to overcome the limitations associated with DE imaging techniques (e.g., increase in exposure dose owing to the requirement of two acquisitions, and emphasis on noise characteristics) via an artificial intelligence technique was presented.ConclusionsThe developed framework addresses X‐ray dose issues in the field of radiation imaging and enables pseudo‐DE imaging with single exposure.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Framework for dual‐energy‐like chest radiography image synthesis from single‐energy computed tomography based on cycle‐consistent generative adversarial network

Lee

et al. 2023

Medical Physics

View full text Add to dashboard Cite

show abstract

“…On [Sirazitdinov et al 2020] was used the ChestX-ray-14, a public dataset provided by [Gusarev et al 2017]. They used 24 images for train 7 to test and 4 to validation, with different models and architecture such as autoencoder, U-Net, cGAN.…”

Section: Related Workmentioning

confidence: 99%

Applying a Conditional GAN for Bone Suppression in Chest Radiography Images

Ziviani

Cámara-Chávez²,

Silva³

2022

Anais Do XLIX Seminário Integrado De Software E Hardware (SEMISH 2022)

View full text Add to dashboard Cite

Bone suppression in radiography is a suitable technique to evaluate the health of soft tissues in exams. For instance, these techniques are essential in evaluating chest radiography images during the COVID-19 outbreak. The purpose of this work is to propose an alternative to solve the bone suppression task in chest radiography images using Generative Adversarial Networks (GANs). Specifically, we used a conditional GAN type (cGAN) to provide a bone-suppressed version of the initial image. To quantify the results, it was necessary to review the main metrics and some state-of-the-art papers related to ours. We compared our result to works from the literature that used the same dataset as the proposal or related techniques. The most used dataset was the Japanese Society of Radiological Technology (JSRT) in these works. With this set of images, we reached a PSNR index of 34.96, which was better than that reviewed in the literature, and a similarity coefficient, known as SSIM, of 0.94. As for the loss calculated by MS-SSIM, we obtained the lowest compared to the reviewed works.

show abstract

Autoencoder-based bone removal algorithm from x-ray images of the lung

Kalisz

Marczyk

2021

2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE)

View full text Add to dashboard Cite

Evaluation of Deep Learning Methods for Bone Suppression from Dual Energy Chest Radiography

Cited by 3 publications

References 14 publications

Framework for dual‐energy‐like chest radiography image synthesis from single‐energy computed tomography based on cycle‐consistent generative adversarial network

Framework for dual‐energy‐like chest radiography image synthesis from single‐energy computed tomography based on cycle‐consistent generative adversarial network

Applying a Conditional GAN for Bone Suppression in Chest Radiography Images

Autoencoder-based bone removal algorithm from x-ray images of the lung

Contact Info

Product

Resources

About