Article
Radiology, Nuclear Medicine & Medical Imaging
Oona Rainio, Chunlei Han, Jarmo Teuho, Sergey V. Nesterov, Vesa Oikonen, Sauli Piirola, Timo Laitinen, Marko Tattalainen, Juhani Knuuti, Riku Klen
Summary: Carimas is a versatile tool for processing medical imaging data, allowing visualization, analysis, and modeling of various medical images in research. Originally designed for positron emission tomography data, it has expanded to include other tomography imaging modalities like computed tomography and magnetic resonance imaging. Carimas excels in analyzing three- and four-dimensional image data and creating polar maps for cardiac perfusion modeling.
JOURNAL OF DIGITAL IMAGING
(2023)
Article
Computer Science, Artificial Intelligence
Juan Wang, Zetao Zhang, Minghu Wu, Yonggang Ye, Sheng Wang, Ye Cao, Hao Yang
Summary: In this work, an improved BlendMask nuclei instance segmentation framework is proposed, which incorporates dilated convolution aggregation module and context information aggregation module to enhance the performance of detecting and segmenting dense small objects and adhering nuclei. A distributional ranking loss function is also introduced to alleviate the imbalance between the target and the background. The proposed method outperforms several recent classic open-source nuclei instance segmentation methods on the DSB2018 dataset, achieving a 3.6% improvement on AP segmentation metric compared to BlendMask.
IET IMAGE PROCESSING
(2023)
Article
Computer Science, Artificial Intelligence
Jhilik Bhattacharya, Tarunpreet Bhatia, Husanbir Singh Pannu
Summary: The paper proposes a search space to narrow down identical images in an archive using Capsule Networks, Wavelet-Discrete Cosine Transform, and Radon Barcodes in medical image retrieval. Empirical case study shows that the method is generic, reproducible, and scalable, improving diagnosis efficiency in automatic image retrieval and annotation.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Yuqing Chen, Zhitao Guo
Summary: In this study, a novel denoising network called TranSpeckle is proposed to remove speckle noise from medical ultrasound images. It utilizes a Transformer block to efficiently capture global contextual relationships without compromising computational efficiency. Additionally, an edge protection module is incorporated to enhance edges, improving the overall despeckling effect of ultrasound images.
IET IMAGE PROCESSING
(2023)
Article
Neurosciences
Hao Guan, Mingxia Liu
Summary: Domain adaptation is an important technique in machine learning-based medical data analysis to reduce distribution differences between different datasets. The Domain Adaptation Toolbox for Medical data analysis (DomainATM) is an open-source software package implemented in MATLAB, offering a collection of popular data adaptation algorithms for medical image analysis. It provides researchers with the capability to perform fast feature-level and image-level adaptation, visualization, and performance evaluation of adaptation methods. Users can also develop and test their own adaptation methods through scripting, enhancing the utility and extensibility of DomainATM. The software, source code, and manual are available online.
Article
Computer Science, Artificial Intelligence
Pengyi Hao, Kangjian Shi, Shuyuan Tian, Fuli Wu
Summary: Medical image segmentation from noisy labels is a challenging task. In this study, a novel uncertainty-aware iterative learning (UaIL) approach is proposed to address issues related to noisy annotations. UaIL trains two deep networks iteratively using original and augmented images, with a joint loss function that includes softened label loss, hard label loss, and consistency loss. Experimental results on two public datasets show that UaIL outperforms competing approaches, especially when dealing with serious label noises. UaIL is also validated on a private dataset, demonstrating its applicability in real-world scenarios with noisy labels.
IET IMAGE PROCESSING
(2023)
Article
Engineering, Biomedical
Xin Wei, Fanghua Ye, Huan Wan, Jianfeng Xu, Weidong Min
Summary: In recent years, deep learning methods have made significant progress in medical image processing. However, the current medical segmentation methods still lack accuracy in segmenting small-scale and variable-scale objects. To address this issue, we propose Triple Attention Network (TANet) which includes a novel Triple Attention Module (TAM). Experimental results demonstrate that TANet outperforms previous models on multiple evaluation metrics and improves the Dice score by up to 7.1%.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Geochemistry & Geophysics
Huachen Yang, Pan Li, Fei Ma, Jianzhong Zhang
Summary: Accurate near-surface velocity models are crucial for land seismic imaging. The conventional first-arrival traveltime tomography (FTT) may fail in complex geological areas. In this study, we propose integrating FTT with supervised deep learning (SDL) to build near-surface velocity models, which significantly reduces the training time and memory costs while achieving satisfactory results.
GEOPHYSICAL JOURNAL INTERNATIONAL
(2023)
Article
Computer Science, Artificial Intelligence
Jin Zhu, Chuan Tan, Junwei Yang, Guang Yang, Pietro Lio
Summary: This paper introduces a novel approach for medical image arbitrary-scale super-resolution, MIASSR, which combines meta-learning and generative adversarial networks (GANs) to achieve super-resolution of medical images at any scale of magnification in (1, 4]. MIASSR shows comparable fidelity performance and the best perceptual quality with the smallest model size compared to state-of-the-art SISR algorithms on single-modal and multi-modal MR brain images. In addition, transfer learning enables MIASSR to handle super-resolution tasks of new medical modalities, with the potential to become a foundational pre-/post-processing step in clinical image analysis tasks.
INTERNATIONAL JOURNAL OF NEURAL SYSTEMS
(2021)
Review
Optics
Kaiyi Tang, Shuangyang Zhang, Zhichao Liang, Yang Wang, Jia Ge, Wufan Chen, Li Qi
Summary: Photoacoustic tomography (PAT) is an imaging technique that uses light-induced acoustic waves for biomedical imaging. Producing high-quality images with PAT is challenging and requires further research. Image post-processing methods have emerged as essential in improving image quality in recent research.
Article
Radiology, Nuclear Medicine & Medical Imaging
Fides R. R. Schwartz, Darin P. P. Clark, Francesca Rigiroli, Kevin Kalisz, Benjamin Wildman-Tobriner, Sarah Thomas, Joshua Wilson, Cristian T. T. Badea, Daniele Marin
Summary: This study evaluated a novel algorithm for noise reduction in obese patients using dual-source dual-energy CT imaging. Seventy-nine patients were included in the retrospective study, comparing the standard clinical algorithm with two new algorithms in terms of image contrast-to-noise ratio, quality, and diagnostic comfort. The results showed that the test algorithm produced higher image quality and was preferred by readers.
EUROPEAN RADIOLOGY
(2023)
Review
Biochemical Research Methods
Vania B. Silva, Danilo Andrade De Jesus, Stefan Klein, Theo van Walsum, Joao Cardoso, Luisa Sanchez Brea, Pedro G. Vaz
Summary: This study provides a comprehensive overview of the methods developed for retrieving speckle information in biomedical OCT applications. The results show that features retrieved from speckle can be successfully used in different applications, such as classification and segmentation. However, the best approach for speckle analysis varies between applications. Further research is needed to validate the applicability and reproducibility of signal-carrying speckle analysis in a clinical context.
JOURNAL OF BIOMEDICAL OPTICS
(2022)
Review
Computer Science, Interdisciplinary Applications
Jaspreet Kaur, Prabhpreet Kaur
Summary: This article discusses the severe impact of COVID-19 outbreak on global health, emphasizing the importance of rapid and accurate diagnosis and control of the epidemic. It explores the application of deep learning techniques in medical imaging and the challenges in responding to health crises.
ARCHIVES OF COMPUTATIONAL METHODS IN ENGINEERING
(2022)
Article
Computer Science, Artificial Intelligence
Jingchen Li, Haobin Shi, Wenbai Chen, Naijun Liu, Kao-Shing Hwang
Summary: In this work, an ensemble-learning-based model with a semi-supervised mechanism is developed for end-to-end medical image detection with overcritical efficiency and accuracy in endoscope detection. A new ensemble mechanism (Al-Adaboost) is proposed to provide a more accurate result through multiple detection models, combining the decision-making of two hierarchical models. Experimental results prove the feasibility and superiority of the model.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Kishore Babu Nampalle, Shriansh Manhas, Balasubramanian Raman
Summary: Since medical images include sensitive patient information, security is crucial during transmission. A new dual encryption method is proposed in this paper, which combines blowfish and signcryption in a certificateless generalized form. The proposed method stands out due to its computational cost-effectiveness and speed. Performance measurements such as PSNR, entropy, MSE, CC, and time taken show that the method is highly secure and efficient.
APPLIED INTELLIGENCE
(2023)