基于机器学习的医学影像分析在药物研发和精准医疗方面的应用

MRI,PET,和CT等医学影像在新药研发和精准医疗中起着越来越重要的作用。影像技术可以被用来诊断疾病,评估药效,选择适应患者,或者确定用药剂量。 随着人工智能技术的发展,特别是机器学习以及深度学习技术在医学影像中的应用,使得我们可以用更短的时间,更少的放射剂量获取更高质量的影像。这些技术还可以帮助放射科医生缩短读片时间,提高诊断准确率。除此之外,机器学习技术还可以提高量化分析的可行性和精度,帮助建立影像与基因以及疾病的临床表现之间的关系。首先根据不同形态的医学影像,简单介绍他们在药物研发和精准医疗中的应用。并对机器学习在医学影像中的功能作一概括总结。最后讨论这个领域的挑战和机遇。     

A systematic review on hyperspectral imaging technology with a machine and deep learning methodology for agricultural applications

The globe's population is increasing day by day, which causes the severe problem of organic food for everyone. Farmers are becoming progressively conscious of the need to control numerous essential factors such as crop health, water or fertilizer use, and harmful diseases in the field. However, it is challenging to monitor agricultural activities. Therefore, precision agriculture is an important decision support system for food production and decision-making. Several methods and approaches have been used to support precision agricultural practices. The present study performs a systematic literature review on hyperspectral imaging technology and the most advanced deep learning and machine learning algorithm used in agriculture applications to extract and synthesize the significant datasets and algorithms. We reviewed legal studies carefully, highlighted hyperspectral datasets, focused on the most methods used for hyperspectral applications in agricultural sectors, and gained insight into the critical problems and challenges in the hyperspectral data processing. According to our study, it has been found that the Hyperion hyperspectral, Landsat-8, and Sentinel 2 multispectral datasets were mainly used for agricultural applications. The most applied machine learning method was support vector machine and random forest. In addition, the deep learning-based Convolutional Neural Networks (CNN) model is mainly used for crop classification due to its high performance with hyperspectral datasets. The present review will be helpful to the new researchers working in the field of hyperspectral remote sensing for agricultural applications with a machine and deep learning methods.     

AI for COVID-19 Detection from Radiographs: Incisive Analysis of State of the Art Techniques,Key Challenges and Future Directions

Background and objectiveIn recent years, Artificial Intelligence has had an evident impact on the way research addresses challenges in different domains. It has proven to be a huge asset, especially in the medical field, allowing for time-efficient and reliable solutions. This research aims to spotlight the impact of deep learning and machine learning models in the detection of COVID-19 from medical images. This is achieved by conducting a review of the state-of-the-art approaches proposed by the recent works in this field.MethodsThe main focus of this study is the recent developments of classification and segmentation approaches to image-based COVID-19 detection. The study reviews 140 research papers published in different academic research databases. These papers have been screened and filtered based on specified criteria, to acquire insights prudent to image-based COVID-19 detection.ResultsThe methods discussed in this review include different types of imaging modality, predominantly X-rays and CT scans. These modalities are used for classification and segmentation tasks as well. This review seeks to categorize and discuss the different deep learning and machine learning architectures employed for these tasks, based on the imaging modality utilized. It also hints at other possible deep learning and machine learning architectures that can be proposed for better results towards COVID-19 detection. Along with that, a detailed overview of the emerging trends and breakthroughs in Artificial Intelligence-based COVID-19 detection has been discussed as well.ConclusionThis work concludes by stipulating the technical and non-technical challenges faced by researchers and illustrates the advantages of image-based COVID-19 detection with Artificial Intelligence techniques.     

Detection and Phenotypic Characterization of Adult Neurogenesis

Studies of adult neurogenesis have greatly expanded in the last decade, largely as a result of improved tools for detecting and quantifying neurogenesis. In this review, we summarize and critically evaluate detection methods for neurogenesis in mammalian and human brain tissue. Besides thymidine analog labeling, cell-cycle markers are discussed, as well as cell stage and lineage commitment markers. Use of these histological tools is critically evaluated in terms of their strengths and limitations, as well as possible artifacts. Finally, we discuss the method of radiocarbon dating for determining cell and tissue turnover in humans.Detection of neurogenesis in vivo requires the ability to image at a cellular resolution, which currently precludes noninvasive imaging approaches, such as magnetic resonance imaging (MRI). In vivo microscopy, using deeply penetrating UV illumination with multiphoton microscopy, or by the recently available endoscopic confocal microscopy, may provide new opportunities for longitudinal studies of neurogenesis in the living animal with single-cell resolution. These newer microscopy approaches are particularly compelling when coupled with transgenic mice expressing phenotype-specific fluorescent reporter genes. Additionally, an advanced method using ¹⁴C carbon dating of postmortem DNA from specific cell populations of the brain revealed insights into adult human neurogenesis. Nevertheless, at present, the predominant approach for studying neurogenesis relies on traditional histological methods of fixation, production of tissue sections, staining, and microscopic analysis.This review discusses methodological considerations for detection of neurogenesis in the adult brain according to our current state of knowledge. This will include the use of exogenous or endogenous markers of cell cycle, as well as phenotype markers that contribute to resolving stages of neuronal lineage commitment. The accurate analysis of cell phenotype will be discussed, including suggestions for accurate detection and reliable quantification of cell numbers. Finally, we will present the newly developed ¹⁴C carbon dating of nuclear DNA for quantitative analysis of neurogenesis in human tissue.     

Improved deep two-photon calcium imaging in vivo

Two-photon laser scanning calcium imaging has emerged as a useful method for the exploration of neural function and structure at the cellular and subcellular level in vivo. The applications range from imaging of subcellular compartments such as dendrites, spines and axonal boutons up to the functional analysis of large neuronal or glial populations. However, the depth penetration is often limited to a few hundred micrometers, corresponding, for example, to the upper cortical layers of the mouse brain. Light scattering and aberrations originating from refractive index inhomogeneties of the tissue are the reasons for these limitations. The depth penetration of two-photon imaging can be enhanced through various approaches, such as the implementation of adaptive optics, the use of three-photon excitation and/or labeling cells with red-shifted genetically encoded fluorescent sensors. However, most of the approaches used so far require the implementation of new instrumentation and/or time consuming staining protocols. Here we present a simple approach that can be readily implemented in combination with standard two-photon microscopes. The method involves an optimized protocol for depth-restricted labeling with the red-shifted fluorescent calcium indicator Cal-590 and benefits from the use of ultra-short laser pulses. The approach allows in vivo functional imaging of neuronal populations with single cell resolution in all six layers of the mouse cortex. We demonstrate that stable recordings in deep cortical layers are not restricted to anesthetized animals but are well feasible in awake, behaving mice. We anticipate that the improved depth penetration will be beneficial for two-photon functional imaging in larger species, such as non-human primates.     

组学时代下机器学习方法在临床决策支持中的应用

随着组学技术的不断发展,对于不同层次和类型的生物数据的获取方法日益成熟。在疾病诊治过程中会产生大量数据,通过机器学习等人工智能方法解析复杂、多维、多尺度的疾病大数据,构建临床决策支持工具,辅助医生寻找快速且有效的疾病诊疗方案是非常必要的。在此过程中,机器学习等人工智能方法的选择显得尤为重要。基于此,本文首先从类型和算法角度对临床决策支持领域中常用的机器学习等方法进行简要综述,分别介绍了支持向量机、逻辑回归、聚类算法、Bagging、随机森林和深度学习,对机器学习等方法在临床决策支持中的应用做了相应总结和分类,并对它们的优势和不足分别进行讨论和阐述,为临床决策支持中机器学习等人工智能方法的选择提供有效参考。     

The European Federation of Organisations for Medical Physics (EFOMP) White Paper: Big data and deep learning in medical imaging and in relation to medical physics profession

Big data and deep learning will profoundly change various areas of professions and research in the future. This will also happen in medicine and medical imaging in particular. As medical physicists, we should pursue beyond the concept of technical quality to extend our methodology and competence towards measuring and optimising the diagnostic value in terms of how it is connected to care outcome. Functional implementation of such methodology requires data processing utilities starting from data collection and management and culminating in the data analysis methods. Data quality control and validation are prerequisites for the deep learning application in order to provide reliable further analysis, classification, interpretation, probabilistic and predictive modelling from the vast heterogeneous big data. Challenges in practical data analytics relate to both horizontal and longitudinal analysis aspects. Quantitative aspects of data validation, quality control, physically meaningful measures, parameter connections and system modelling for the future artificial intelligence (AI) methods are positioned firmly in the field of Medical Physics profession. It is our interest to ensure that our professional education, continuous training and competence will follow this significant global development.     

Artificial intelligence and machine learning for medical imaging: A technology review

Artificial intelligence (AI) has recently become a very popular buzzword, as a consequence of disruptive technical advances and impressive experimental results, notably in the field of image analysis and processing. In medicine, specialties where images are central, like radiology, pathology or oncology, have seized the opportunity and considerable efforts in research and development have been deployed to transfer the potential of AI to clinical applications. With AI becoming a more mainstream tool for typical medical imaging analysis tasks, such as diagnosis, segmentation, or classification, the key for a safe and efficient use of clinical AI applications relies, in part, on informed practitioners. The aim of this review is to present the basic technological pillars of AI, together with the state-of-the-art machine learning methods and their application to medical imaging. In addition, we discuss the new trends and future research directions. This will help the reader to understand how AI methods are now becoming an ubiquitous tool in any medical image analysis workflow and pave the way for the clinical implementation of AI-based solutions.     

An experiment on animal re-identification from video

In the face of the global concern about climate change and endangered ecosystems, monitoring individual animals is of paramount importance. Computer vision methods for animal recognition and re-identification from video or image collections are a modern alternative to more traditional but intrusive methods such as tagging or branding. While there are many studies reporting results on various animal re-identification databases, there is a notable lack of comparative studies between different classification methods. In this paper we offer a comparison of 25 classification methods including linear, non-linear and ensemble models, as well as deep learning networks. Since the animal databases are vastly different in characteristics and difficulty, we propose an experimental protocol that can be applied to a chosen data collections. We use a publicly available database of five video clips, each containing multiple identities (9 to 27), where the animals are typically present as a group in each video frame. Our experiment involves five data representations: colour, shape, texture, and two feature spaces extracted by deep learning. In our experiments, simpler models (linear classifiers) and just colour feature space gave the best classification accuracy, demonstrating the importance of running a comparative study before resorting to complex, time-consuming, and potentially less robust methods.     

基于机器学习的风景园林智能化分析应用研究

机器学习使实现数据的智能化处理及充分利用数据中蕴含的知识与价值成为可能。探索基于机器学习在风景园林领域智能化分析应用的途径,开展3个实验。其中2个与数据分析研究相关,提出基于调研图像色彩聚类分析的城市色彩印象和基于图像识别技术的景观视觉质量评估与网络应用平台部署实验。最后1个实验与数字化设计创作相关,提出用于设计方案遴选的地形生成方法,包括2个子项目：应用深度学习生成对抗网络（GAN）的地形生成和建立遮罩、预测未知区域的高程。3个实验应用到机器学习中分类、聚类和回归3个主要方向中的算法以及深度学习的生成对抗网络,对传统的研究问题提出了基于机器学习新的研究方法。因此,在应用机器学习风景园林领域,可以有效地从多源数据中学习相互增强的知识,发现问题,并提出解决问题的新方法。     

Imaging for lung physiology: what do we wish we could measure?

The role of imaging as a tool for investigating lung physiology is growing at an accelerating pace. Looking forward, we wished to identify unresolved issues in lung physiology that might realistically be addressed by imaging methods in development or imaging approaches that could be considered. The role of imaging is framed in terms of the importance of good spatial and temporal resolution and the types of questions that could be addressed as these technical capabilities improve. Recognizing that physiology is fundamentally a quantitative science, a recurring emphasis is on the need for imaging methods that provide reliable measurements of specific physiological parameters. The topics included necessarily reflect our perspective on what are interesting questions and are not meant to be a comprehensive review. Nevertheless, we hope that this essay will be a spur to physiologists to think about how imaging could usefully be applied in their research and to physical scientists developing new imaging methods to attack challenging questions imaging could potentially answer.     

Super‐beacons: Open‐source probes with spontaneous tuneable blinking compatible with live‐cell super‐resolution microscopy

Localization‐based super‐resolution microscopy relies on the detection of individual molecules cycling between fluorescent and non‐fluorescent states. These transitions are commonly regulated by high‐intensity illumination, imposing constrains to imaging hardware and producing sample photodamage. Here, we propose single‐molecule self‐quenching as a mechanism to generate spontaneous photoswitching. To demonstrate this principle, we developed a new class of DNA‐based open‐source super‐resolution probes named super‐beacons, with photoswitching kinetics that can be tuned structurally, thermally and chemically. The potential of these probes for live‐cell compatible super‐resolution microscopy without high‐illumination or toxic imaging buffers is revealed by imaging interferon inducible transmembrane proteins (IFITMs) at sub‐100 nm resolutions.     

AI applications to medical images: From machine learning to deep learning

PurposeArtificial intelligence (AI) models are playing an increasing role in biomedical research and healthcare services. This review focuses on challenges points to be clarified about how to develop AI applications as clinical decision support systems in the real-world context.MethodsA narrative review has been performed including a critical assessment of articles published between 1989 and 2021 that guided challenging sections.ResultsWe first illustrate the architectural characteristics of machine learning (ML)/radiomics and deep learning (DL) approaches. For ML/radiomics, the phases of feature selection and of training, validation, and testing are described. DL models are presented as multi-layered artificial/convolutional neural networks, allowing us to directly process images. The data curation section includes technical steps such as image labelling, image annotation (with segmentation as a crucial step in radiomics), data harmonization (enabling compensation for differences in imaging protocols that typically generate noise in non-AI imaging studies) and federated learning. Thereafter, we dedicate specific sections to: sample size calculation, considering multiple testing in AI approaches; procedures for data augmentation to work with limited and unbalanced datasets; and the interpretability of AI models (the so-called black box issue). Pros and cons for choosing ML versus DL to implement AI applications to medical imaging are finally presented in a synoptic way.ConclusionsBiomedicine and healthcare systems are one of the most important fields for AI applications and medical imaging is probably the most suitable and promising domain. Clarification of specific challenging points facilitates the development of such systems and their translation to clinical practice.     

Imaging and characterisation of the surface of live cells

Determining the organisation of key molecules on the surface of live cells in two dimensions and how this changes during biological processes, such as signaling, is a major challenge in cell biology and requires methods with nanoscale resolution. Recent advances in fluorescence imaging both at the diffraction limit tracking single molecules and exploiting super resolution imaging have now reached a stage where they can provide fundamentally new insights. Complementary developments in scanning ion conductance microscopy also allow the cell surface to be imaged with nanoscale resolution. The challenge now is to combine the information obtained using these different methods and on different cells to obtain a coherent view of the cell surface. In the future this needs to be driven by interdisciplinary research between physical scientists and biologists.     

Systematic comparison of SCOP and CATH: a new gold standard for protein structure analysis

Background  

SCOP and CATH are widely used as gold standards to benchmark novel protein structure comparison methods as well as to train machine learning approaches for protein structure classification and prediction. The two hierarchies result from different protocols which may result in differing classifications of the same protein. Ignoring such differences leads to problems when being used to train or benchmark automatic structure classification methods. Here, we propose a method to compare SCOP and CATH in detail and discuss possible applications of this analysis.     

X10 expansion microscopy enables 25‐nm resolution on conventional microscopes

Expansion microscopy is a recently introduced imaging technique that achieves super‐resolution through physically expanding the specimen by ~4×, after embedding into a swellable gel. The resolution attained is, correspondingly, approximately fourfold better than the diffraction limit, or ~70 nm. This is a major improvement over conventional microscopy, but still lags behind modern STED or STORM setups, whose resolution can reach 20–30 nm. We addressed this issue here by introducing an improved gel recipe that enables an expansion factor of ~10× in each dimension, which corresponds to an expansion of the sample volume by more than 1,000‐fold. Our protocol, which we termed X10 microscopy, achieves a resolution of 25–30 nm on conventional epifluorescence microscopes. X10 provides multi‐color images similar or even superior to those produced with more challenging methods, such as STED, STORM, and iterative expansion microscopy (iExM). X10 is therefore the cheapest and easiest option for high‐quality super‐resolution imaging currently available. X10 should be usable in any laboratory, irrespective of the machinery owned or of the technical knowledge.     

Efficient super‐resolution volumetric imaging by radial fluctuation Bayesian analysis light‐sheet microscopy

Various computational super‐resolution methods are available based on the analysis of fluorescence fluctuation behind acquired frames. However, dilemmas often exist in the balance of fluorophore characteristics, computation cost, and achievable resolution. Here we present an approach that uses a super‐resolution radial fluctuations (SRRF) image to guide the Bayesian analysis of fluorophore blinking and bleaching (3B) events, allowing greatly accelerated localization of overlapping fluorophores with high accuracy. This radial fluctuation Bayesian analysis (RFBA) approach is also extended to three dimensions for the first time and combined with light‐sheet fluorescence microscopy, to achieve super‐resolution volumetric imaging of thick samples densely labeled with common fluorophores. For example, a 700‐nm thin Bessel plane illumination is developed to optically section the Drosophila brain, providing a high‐contrast 3D image of rhythmic neurons. RFBA analyzes 30 serial volumes to reconstruct a super‐resolved 3D image at 4‐times higher resolutions (~70 and 170 nm), and precisely resolve the axon terminals. The computation is over 2‐orders faster than conventional 3B analysis microscopy. The capability of RFBA is also verified through dual‐color imaging of cell nucleus in live Drosophila brain. The spatial co‐localization patterns of the nuclear envelope and DNA in a neuron deep inside the brain can be precisely extracted by our approach.     

Emerging deep learning methods for single-cell RNA-seq data analysis

Deep learning is making major breakthrough in several areas of bioinformatics. Anticipating that this will occur soon for the single-cell RNA-seq data analysis, we review newly published deep learning methods that help tackle computational challenges. Autoencoders are found to be the dominant approach. However, methods based on deep generative models such as generative adversarial networks (GANs) are also emerging in this area.     

NMR studies on heavy metal immobilization in biosorbents and mineral matrices

Magnetic resonance imaging and otherNMR methods have a potential for thenon-destructive observation of environmentallyrelevant processes with both spatial andtemporal resolution. Among other applications,such methods can be used to study transport andimmobilization of paramagnetic heavy metal ionsin biosorbents and other matrices. Thisoverview covers various NMR approaches to studysuch processes and illustrates them withexamples of imaging on alginate-basedbiosorbents and on heavy-metal doped gypsumpastes. Experimental challenges in studies ofother matrices are shortly addressed as well.