In recent years, multimodal medical image fusion (MMIF) has attracted significant attention due to its ability to integrate complementary information from different medical imaging modalities and provide more comprehensive information for clinical analysis. By combining anatomical information from modalities such as computed tomography (CT) and magnetic resonance imaging (MRI) with functional information from positron emission tomography (PET) and single-photon emission computed tomography (SPECT), MMIF can improve image quality and support subsequent tasks such as disease analysis, lesion detection, image segmentation, and treatment planning. This review provides a comprehensive overview of MMIF from theoretical and technical perspectives. First, commonly used medical imaging modalities and publicly available medical image databases are summarized and compared. Subsequently, the general workflow, fusion levels, and quality requirements of MMIF are introduced. Representative fusion techniques are then systematically reviewed, including spatial-domain methods, transform-domain methods, sparse representation-based methods, deep learning-based methods, hybrid methods, and emerging Mamba-based approaches. In addition, commonly used image fusion quality assessment metrics are analyzed, and the reported quantitative performance of representative MMIF methods is compared and discussed. Finally, current challenges and future development trends of MMIF are presented, including robustness, clinical translation, and emerging multimodal learning paradigms.

In terms of speed and accuracy, the deep learning-based polyp segmentation method is superior. It is essential for the early detection and treatment of colorectal cancer and has the potential to greatly reduce the disease's overall prevalence. Due to the various forms and sizes of polyps, as well as the blurring of the boundaries between the polyp region and the surrounding mucus, most existing algorithms are unable to provide highly accurate colorectal polyp segmentation. Therefore, to overcome these obstacles, we propose an adaptive feature aggregation network (AFANet). It contains two main modules: the Multi-modal Balancing Attention Module (MMBA) and the Global Context Module (GCM). The MMBA extracts improved local characteristics for inference by integrating local contextual information while paying attention to them in three regions: foreground, background, and border. The GCM takes global information from the top of the encoder and sends it to the decoder layer in order to further investigate global contextual feature information in the pathologic picture. Dice of 92.11 % and 94.76 % and MIoU of 91.07 % and 94.54 %, respectively, are achieved by comprehensive experimental validation of our proposed technique on two benchmark datasets, Kvasir-SEG and CVCClinicDB. The experimental results demonstrate that the strategy outperforms other cutting-edge approaches.

The automatic segmentation of diabetic retinopathy (DR) holds significant importance for assisting physicians in diagnosis and treatment. Given the complexity, high inter-class similarity, and uncertainty of DR, it is crucial to integrate multiscale information between lesions and establish global correlations among them. To address these issues, a novel HRU-TNet (Hybrid Residual U-Transformer Network) algorithm for retinal lesion segmentation is proposed. In this framework, the network is augmented with lightweight self-attention residual U-modules (LSA-RSU) to capture high-frequency details of the lesions and global contextual information. The skip connections are then enhanced through interactive residual transformer fusion modules (IRTF) and channel-cross attention (CCA), promoting dependencies among features at different scales and filtering out interfering information to guide feature fusion and eliminate ambiguity. Additionally, a novel retinal image enhancement technique is devised, employing local wavelet transformations to capture detailed components of the retinal images, thereby enhancing the representational capacity of the segmentation network. Data augmentation is also performed to ensure network adaptability to small datasets. Comprehensive experiments conducted on the publicly available IDRID and e_ophtha datasets yielded average AUC_PR values of 0.709 and 0.451, respectively. The proposed approach demonstrated superior generalization on the DDR dataset compared to other methods mentioned in the literature. These results demonstrate that our proposed method is better suited for small retinal datasets, exhibiting improved segmentation accuracy and generalization compared to existing approaches.

针对主流目标检测算法在检测SAR图像舰船小目标时精度低、实时性能差等问题,本文提出了一种基于改进YOLOV4-Tiny的目标检测算法.根据SAR图像分辨率低,特征简单且含有大量小目标的特点,首先,增加8倍下采样特征图来增强对小目标特征的提取能力;其次,借鉴Inception和SPP网络设计出新的特征提取模块Module来获取更高级的语义信息且提高浅层网络的感受野,进一步提升检测小目标的精度,耗时仅增加1ms.引入SE通道注意力机制来增强重要特征,提高网络的鲁棒性,并通过对比实验选出SE通道注意力机制最合适的参数.实验结果表明,改进后YOLOV4-Tiny的AP值达到97.17％,单张图像检测耗时0.022s,参数量为9.22M,AP值相对原始模型提高2.93％,检测耗时增加8ms,参数量增加3.35M,满足实时性检测需求.

Deep learning techniques are frequently employed in the field of medical image segmentation, its speed and accuracy advantages over more conventional methods. Colorectal polyp segmentation technology is one such tool, which is essential in the early detection and treatment of colorectal cancer and has the potential to greatly reduce the disease's overall prevalence. Due to the polyps’ varied shapes,sizes,colors,as well as the colonoscopic images’ low contrast and blurring boundaries between the polyp region and the surrounding mucus,most existing algorithms fail to achieve highly accurate colorectal polyp segmentation.To overcome these obstacles, we present an adaptive context-selection network that can accurately identify polyps from colonoscopy pictures by utilizing the attention balancing module and channel attention. In particular, the network is an encoder-decoder structure wherein local and global contextual features are refined during the encoding stage by means of channel attention (Squeeze-and-Excite,SE). To improve local feature extraction, the BAM uses attention mechanisms in the background, foreground, and boundary. Similarity coefficients of 92.11% and 94.76%, and average cross-merge ratios of 91.07% and 94.54%, respectively, are achieved by comprehensive experimental validation of our proposed technique on two benchmark datasets, Kvasir-SEG and CVC-ClinicDB. The experimental results demonstrate the strategy outperforms other cutting-edge approaches.

为解决图像配准中因旋转变换,尺度缩放,光照变化,相机抖动,气流变化以及压缩等影响成像质量,而导致特征点检测数目不足的问题,本文提出了基于组合模型的图像配准方法.该方法采用KAZE与SURF算子联合检测局部区域中线性与非线性特征点;同时为了提高计算效率,采用二进制向量描述符表征特征点,并使用汉明距离计算特征点之间的匹配距离,有效地提高匹配效率;最后,利用随机一致性算法(RANSAC)进一步消除异常点,并根据内点之间的对应关系来计算几何变换模型.实验结果表明:本文有效地解决了因特征点数目不足配准失效的问题,多幅图像的配准实验结果说明了本文方法具有更好的稳定性与鲁棒性,同时运行效率最快.

For image registration, feature detection and description are critical steps that identify the keypoints and describe them for the subsequent matching to estimate the geometric transformation parameters between two images. Recently, there has been a large increase in the research methods of detection operators and description operators, from traditional methods to deep learning methods. To solve the problem, that is, which operator is suitable for specific application problems under different imaging conditions, the paper systematically reviewed commonly used descriptors and detectors from artificial methods to deep learning methods, and the corresponding principle, analysis, and comparative experiments are given as well. We introduce the handcrafted detectors including FAST, BRISK, ORB, SURF, SIFT, and KAZE and the handcrafted descriptors including BRISK, FREAK, BRIEF, SURF, ORB, SIFT, KAZE. At the same time, we review detectors based on deep learning technology including DetNet, TILDE, LIFT, multiscale detector, SuperPoint, and descriptors based on deep learning including pretrained descriptor, Siamese descriptor, LIFT, triplet network, and SuperPoint. Two group of comparison experiments are compared comprehensively and objectively on representative datasets. Finally, we concluded with insightful discussions and conclusions of descriptor and detector selection for specific application problem and hope this survey can be a reference for researchers and engineers in image registration and related fields.

The traditional principle of synthetic aperture radio imaging is based on Shannon sampling theorem, which obtain the complete spectrum data with Shannon sampling and the inverse Fourier transform is used to generate the image. Due to the imaging equipment and external environmental factors, the spectrum data is accompanied by a large number of unreal signal, which causes a large amount of noise to be generated in the image, usually called a dirty image. In the field of radio astronomy, the related clean algorithm is usually used to process the dirty image to obtain a "clean" image. In order to reduce the sampling cost of the radio signal and obtain more "clean" radio image, based on radio interference sparse imaging and compressed sensing theory, we realize the reconstruction of the dirty image from the incomplete spectrum, and then removes the noise, orthogonal matching pursuit and feature sign algorithm is used to reconstruct the dirty image from the sparse spectrum and the noise is then removed by block-matching and 3-D filtering.

Image enhancement is a challenging problem in the field of image processing, especially low-light color images enhancement. This paper proposed a robust and comprehensive enhancement method based several points. First, the idea of bright channel is introduced to estimate the illumination map which is used to attain the enhancing result with Retinex model, and the color constancy is keep as well. Second, in order eliminate the illumination offsets wrongly estimated, morphological closing operation is used to modify the initial estimating illumination. Furthermore, in order to avoid fabricating edges, enlarged noises and over-smoothed visual features appearing in enhancing result, a multi-scale closing operation is used. At last, in order to avoiding the haloes and artifacts presented in enhancing result caused by gradient information lost in previous step, guided filtering is introduced to deal with previous result with guided image is initial bright channel. The proposed method can get good illumination map, and attain very effective enhancing results, including dark area is enhanced with more visual features, color natural and constancy, avoiding artifacts and over-enhanced, and eliminating Incorrect light offsets.

The optimization inspired network can bridge convex optimization and neural networks in Compressive Sensing (CS) reconstruction of natural image, like ISTA-Net+, which mapping optimization algorithm: iterative shrinkage-thresholding algorithm (ISTA) into network. However, measurement matrix and input initialization are still hand-crafted, and multi-channel feature map contain information at different frequencies, which is treated equally across channels, hindering the ability of CS reconstruction in optimization-inspired networks. In order to solve the above problems, we proposed MC-ISTA-Net

针对复杂产品装配设计过程中零部件间匹配关系描述不清晰,装配设计过程中零件配对精度不高等问题,提出了可装配特征的结构化描述与匹配推理研究方法.在分析影响零件获取、匹配、调整以及装配的工艺关联、匹配型面特征,以及装配设计意图的基础上,给出了包括工程语义、装配约束、装配端口、装配空间关系的装配关键结的定义和结构化描述.基于装配关键结的复杂产品零部件进行了可装配性分析,确定了装配关键结中各因素权重,并通过模糊理想解法完成了对复杂零部件装配属性邻接图的求解,最终通过复杂产品装配从定性输入到定量求解,再到定性输出的映射变换实现了零部件间的精准匹配,给出了推理匹配过程.最后,通过应用算例验证了方法的有效性.

It is urgent to combine knowledge resources with manufacturing business processes to form a knowledge service in the cloud mode, so as to provide intelligent support for business activities in product development process. The main challenge of knowledge resource service, however, is how to rapidly construct the complex resource service system and respond promptly to the changeable service requirements in the business process, which is similar to the software system modeling using a component in software engineering. This paper is concerned with an optimal composition framework (OCF) of knowledge resource service, including service decomposition, component encapsulation, and optimal composition. Firstly, the typical business processes are decomposed into the dynamic knowledge element (DKE), and all kinds of knowledge resources and service behaviors are encapsulated into the reusable resource service components (RSC). Then, a multicomponent optimal composition mathematical model is presented, which transforms the problem of component composition into a multiobjective optimization problem. On this basis, a heuristic algorithm with the adaptive mutation probability is introduced to composite the multigranularity service component dynamically and robustly. Finally, the case of component composition for maintenance resource service is studied and the simulation results are provided to verify the efficacy of the proposed model and algorithms.

The high resolution image reconstruction takes an important place in the solar physics research,but the solar high resolution observation has been hindered severely for a long time due to huge observation data volume,slow reconstruction speed and other factors.In order to treat the huge volume of quasi real-time solar observation data and cope with the computing burden of the same magnitude for high resolution reconstruction of solar image,a number of advanced ground-based solar telescopes at home and abroad have adopted the speckle masking,a method of reconstruction algorithm that can be parallely realized,to reconstruct the high resolution image.Good treatment results are obtained from this method.However,it is still hard to meet the demand for solar observation data treatment at the current efficiency,since the volume of solar telescope observation data is increasing.This paper is directed at the demands of China's new solar telescopes such as the NVST (The 1 m New Vacuum Solar Telescope) and ONSET (Optical and Near-Infrared Solar Eruption Tracer).By analyzing the computing time of each module with the Triple-Spectral method through actual measurement,the paper reaches a conclusion that the data exchange performance is a key bottleneck affecting the reconstruction effect.On this basis,the paper puts forward a universal method of global-shared low-exchange parallel high resolution image reconstruction.This method takes the Triple-Spectral as the core algorithm for image reconstruction and uses the message passing interface (MPI) and shared memory mechanism,allowing the reconstruction computing process to read and write the data in the shared memory at high speed after algorithm optimization.The shared memory,which is created for storage of image data and image reconstruction results respectively according to the size of the image,will be subsequently mapped into each process,giving access to read the processing data and store the reconstruction results independently.While computing the image reconstruction,each child process will not apply the MPI communication to obtain the sub-block image data.Instead,it will read the associated data from the shared memory according to the sub-block image numbering.After the sub-block image reconstruction is done,each child process will store the reconstruction results directly into the shared memory according to the sub-block image numbering,instead of sending the results to the host process via MPI communication.In this way the communication between processes is reduced,the comunication time saved and the data exchange efficiency improved.The experiment results show that on one 16-core PC server,it takes only about 12.4 s for reconstructing the 100-frame ONSET 1660×1660 pixel image and 5.6 s for the 100-frame NVST 1024×1024 pixel image under this method.Good efficiency is achieved in the reconstruction of solar telescope data for ONSET and NVST with different apertures,proving that this method has certain efficiency and universality.The parallel combination of the Triple-Spectral and K-T has greatly reduced the communication process and data exchange in the course of image reconstruction,saved the communication time and improved the reconstruction efficiency.To be further mentioned,the achievement of good efficiency on one server would bring the method into flexible application and save the equipment cost substantially.With the aid of the research outcome of the paper,it is expected to tackle the puzzles remained in the high resolution reconstruction of NVST and ONSET,and bring down the data storage burden.Its fulfillment of real-time high resolution reconstruction has laid a solid foundation for the follow-up researches.

As a dedicated solar radio interferometer, the MingantU SpEctral RadioHeliograph (MUSER) generates massive observational data in the frequency range of 400 MHz-15 GHz. High-performance imaging forms a significantly important aspect of MUSER's massive data processing requirements. In this study, we implement a practical high-performance imaging pipeline for MUSER data processing. At first, the specifications of the MUSER are introduced and its imaging requirements are analyzed. Referring to the most commonly used radio astronomy software such as CASA and MIRIAD, we then implement a high-performance imaging pipeline based on the Graphics Processing Unit technology with respect to the current operational status of the MUSER. A series of critical algorithms and their pseudo codes, i.e., detection of the solar disk and sky brightness, automatic centering of the solar disk and estimation of the number of iterations for clean algorithms, are proposed in detail. The preliminary experimental results indicate that the proposed imaging approach significantly increases the processing performance of MUSER and generates images with high-quality, which can meet the requirements of the MUSER data processing.

Kafka is a high-throughput distributed messaging system.It is widely used in the distributed massive data proces-sing and other fields.But the current load balancing algorithm for the Kafka consumer client has some shortcomings.It will increase the overhead and even make some mistakes in the massive data processing.This paper developed an optimized load balancing algorithm for the Kafka consumer client.The process of load balancing was totally controlled by the consumer which was the manager.The rest of consumers didn''t need to do load balancing alone.And the manager didn''t have to redistribute partitions among consumers.The system monitored the condition of every consumer and did load balancing in time if there was a crashed consumer.The results of tests show the algorithm can reduce the system overhead caused by load balancing and avoid getting wrong results of load balancing.It can guarantee the correctness of distributed scientific data processing effectively.

Astronomical data processing is a very important part of astronomical research.With the rapid growth of function and observation capability of the new generation of telescopes,building a high performance real-time computing platform in observing sites and handling the data quickly have become a trend.Based on the real-time data processing system construction requirement of MUSER,this paper has systematically studied the implementation method based upon OpenStack local cloud and automation management pattern.We propose a strategy which can open or close compute nodes automatically and the experiments are carried out.Experiments show that this pattern can meet the needs of astronomical data processing completely and it is more efficient than other traditional data processing methods of static computing resources allocation.At the same time,it can efficiently save our energy bills and reduce the observing costs.This pattern has certain value for reference to build a high-performance computing platform for observatories in the future.

Astronomical remote monitoring is an important part of astronomical research.To detect the distortion data caused by weather condition,electromagnetic interference and antenna malfunction in time and improve the reliability of data processing,it is the urgent demand for MUSER (Mingantu Ultrawide Spectral Radio Heliograph,the original name is Chinese Spectral Radio Heliograph-CSRH) to realize the remote monitoring of the state of real time observation.In this paper,we present a visualization method for MUSER remote monitoring data which is based on WebSocket technology.And we use power and spectrum data in real-time visualization display as an example to validate.The experimental results show that proposed method can meet the requirements of real-time observation monitoring of MUSER.It changes the traditional local monitor mode,breaks the region limitation and has series of advantages such as real-time response,maneuverability and extensibility.At the same times,the proposed method improves the efficiency and equipment utilization.And there is a certain reference value to other remote monitoring and data visualization systems.

As a dedicated synthetic aperture radio interferometer in China, the MingantU SpEctral Radioheliograph (MUSER), initially known as the Chinese Spectral RadioHeliograph (CSRH), has entered the stage of routine observation. More than 23 million data records per day need to be effectively managed to provide high-performance data query and retrieval for scientific data reduction. In light of these massive amounts of data generated by the MUSER, in this paper, a novel data management technique called the negative database (ND) is proposed and used to implement a data management system for the MUSER. Based on the key-value database, the ND technique makes complete utilization of the complement set of observational data to derive the requisite information. Experimental results showed that the proposed ND can significantly reduce storage volume in comparison with a relational database management system (RDBMS). Even when considering the time needed to derive records that were absent, its overall performance, including querying and deriving the data of the ND, is comparable with that of a relational database management system (RDBMS). The ND technique effectively solves the problem of massive data storage for the MUSER and is a valuable reference for the massive data management required in next-generation telescopes.

The volume of data generated by modern astronomical telescopes is extremely large and rapidly growing. However, current high-performance data processing architectures/frameworks are not well suited for astronomers because of their limitations and programming difficulties. In this paper, we therefore present OpenCluster, an open-source distributed computing framework to support rapidly developing high-performance processing pipelines of astronomical big data. We first detail the OpenCluster design principles and implementations and present the APIs facilitated by the framework. We then demonstrate a case in which OpenCluster is used to resolve complex data processing problems for developing a pipeline for the Mingantu Ultrawide Spectral Radioheliograph. Finally, we present our OpenCluster performance evaluation. Overall, OpenCluster provides not only high fault tolerance and simple programming interfaces, but also a flexible means of scaling up the number of interacting entities. OpenCluster thereby provides an easily integrated distributed computing framework for quickly developing a high-performance data processing system of astronomical telescopes and for significantly reducing software development expenses.

In the process of a distributed high￣speed real-time storage of data from a large solar telescope, inconsistency occurs between FITS header, a kind of meta-data, and FITS data files, which causes applications unable to retrieve correct FITS data files when applying the index information built on the meta-data. In this paper, models of consistency between the above-mentioned two types of data are studied during high-speed data storage;and the mechanism and solutions are summarized to achieve consistency for real-time data storage. A two-phase commit protocol ( 2PC ) is proposed to ensure consistency, and 2PC algorithm implementation in high-speed data storage is given. The present study on different applications of data storage, with its unique research process and methods, can serve as references to other fields of scientific study.