paper_authors: Gangwon Jeong, Fu Li, Umberto Villa, Mark A. Anastasio
for: 这个研究的目的是为了使用深度学习方法来提高ultrasound computed tomography（USCT）中的速度声速（SOS）的重建精度，并且 investigate the impact of chosen input modalities on image-to-image learned reconstruction（IILR）方法。
results: 研究发现，使用双通道输入可以提高IILR方法的重建精度和特征区域的检测性能。单通道输入（TT或RT图像）alone的情况下，重建精度和特征区域的检测性能均较差。Abstract
Ultrasound computed tomography (USCT) is actively being developed to quantify acoustic tissue properties such as the speed-of-sound (SOS). Although full-waveform inversion (FWI) is an effective method for accurate SOS reconstruction, it can be computationally challenging for large-scale problems. Deep learning-based image-to-image learned reconstruction (IILR) methods are being investigated as scalable and computationally efficient alternatives. This study investigates the impact of the chosen input modalities on IILR methods for high-resolution SOS reconstruction in USCT. The selected modalities are traveltime tomography (TT) and reflection tomography (RT), which produce a low-resolution SOS map and a reflectivity map, respectively. These modalities have been chosen for their lower computational cost relative to FWI and their capacity to provide complementary information: TT offers a direct -- while low resolution -- SOS measure, while RT reveals tissue boundary information. Systematic analyses were facilitated by employing a stylized USCT imaging system with anatomically realistic numerical breast phantoms. Within this testbed, a supervised convolutional neural network (CNN) was trained to map dual-channel (TT and RT images) to a high-resolution SOS map. Moreover, the CNN was fine-tuned using a weighted reconstruction loss that prioritized tumor regions to address tumor underrepresentation in the training dataset. To understand the benefits of employing dual-channel inputs, single-input CNNs were trained separately using inputs from each modality alone (TT or RT). The methods were assessed quantitatively using normalized root mean squared error and structural similarity index measure for reconstruction accuracy and receiver operating characteristic analysis to assess signal detection-based performance measures.
摘要
美国计算 Tomography (USCT) 目前在发展中，以量化声学组织特性，如声速 (SOS) 为目标。虽然全波形反射 (FWI) 是一种高精度的 SOS 重建方法，但可能会对大规模问题具有计算挑战。深度学习基于图像到图像学习的方法被调查为可扩展和计算高效的替代方案。本研究研究了选择的输入模式对 IILR 方法的高分辨率 SOS 重建影响。选择的模式包括旅游时间 Tomography (TT) 和反射 Tomography (RT)，它们生成了低分辨率 SOS 地图和反射图像，分别。这些模式选择的原因是它们的计算成本较低，并且可以提供补充信息：TT 提供了直接 -- 低分辨率 -- SOS 测量，而 RT 揭示了组织边界信息。在使用静态 USCT 图像系统和数字胸部phantom进行系统性分析的测试环境中，一个以图像为输入的 convolutional neural network (CNN) 被训练来将双通道 (TT 和 RT 图像) 映射到高分辨率 SOS 地图。此外，CNN 还被微调使用一个权重重建损失函数，以优先级刻意诊断区域，以 Addressing tumor underrepresentation in the training dataset。为了了解使用双通道输入的好处，单通道 CNNs 分别使用每个模式的输入图像来训练（TT 或 RT）。这些方法被评估量化使用 normalized root mean squared error 和 structure similarity index measure 来评估重建精度和 Receiver operating characteristic analysis 来评估基于信号检测的性能指标。
Combined Channel and Spatial Attention-based Stereo Endoscopic Image Super-Resolution
results: 根据da Vinci数据集的训练，提出的模型可以提高PSNR值达2.12 dB（比较2）和1.29 dB（比较4），同时SSIM值也提高了0.03（比较2）和0.0008（比较4）。这种方法可以帮助医生和Physician更准确地诊断和治疗endoscopic图像。Abstract
Stereo Imaging technology integration into medical diagnostics and surgeries brings a great revolution in the field of medical sciences. Now, surgeons and physicians have better insight into the anatomy of patients' organs. Like other technologies, stereo cameras have limitations, e.g., low resolution (LR) and blurry output images. Currently, most of the proposed techniques for super-resolution focus on developing complex blocks and complicated loss functions, which cause high system complexity. We proposed a combined channel and spatial attention block to extract features incorporated with a specific but very strong parallax attention module (PAM) for endoscopic image super-resolution. The proposed model is trained using the da Vinci dataset on scales 2 and 4. Our proposed model has improved PSNR up to 2.12 dB for scale 2 and 1.29 dB for scale 4, while SSIM is improved by 0.03 for scale 2 and 0.0008 for scale 4. By incorporating this method, diagnosis and treatment for endoscopic images can be more accurate and effective.
摘要
单声图像技术在医疗诊断和手术中得到了很大的革命，为医疗科学带来了更好的顾问。现在医生和医生都可以更好地了解患者的器官 анато�。然而，单声摄像头也有其限制，例如低分辨率（LR）和模糊的输出图像。现在大多数提议的超解析技术都是建立复杂的封包和复杂的损失函数，这会导致高系统复杂性。我们提出了一个结合通道和空间注意力块的特殊专注模组（PAM），用于检测照片超解析。我们的提议模型在 scales 2 和 4 上训练，实现了 PSNR 的提升至 2.12 dB 和 1.29 dB，而 SSIM 则提高了 0.03 和 0.0008。通过这种方法，医疗诊断和治疗可以更精准和有效。