Deep reinforcement learning-based user scheduling methods with low-complexity beamforming for massive MU-MIMO systems

Image classification is a fundamental problem in the field of computer vision and involves assigning a label to an image based on its content. In this paper, we survey both traditional machine learning-based image classification methods and deep learning-based image classification methods, specifically, we review two deep learning-based image classification methods: a convolutional neural network (CNN) method and a pre-trained CNN-based transfer learning method for image classification. We first briefly review traditional machine learning-based image classification methods and then deep learning-based image classification methods, discussing both the feature extraction methods and the classification methods used in the deep learning-based methods. We discuss several deep neural network architectures for image classification, including LeNet, AlexNet, VGGNet, GoogleNet, ResNet, and DenseNet, and finally, we conclude with a discussion on the applications of image classification and compare the different methods based on various factors both qualitatively and quantitatively, and present our experimental results, while also summarizing various methods of data augmentation, batch normalization, and regularization by dropout. We introduce techniques for practical training of deep networks, and discuss fine-tuning, pruning, and model quantization for efficient inference of trained neural networks, and finally, we introduce and discuss the runtime engines for the deployment of trained neural networks for efficient inference.

Illumination estimation is a fundamental prerequisite for many computer vision applications. Various statistics and deep learning-based estimation methods have been proposed, and further studies are ongoing. In this study, we first perform a comparative analysis of representative statistics and deep learning-based methods and subsequently investigate combining them to improve the illumination estimation accuracy. We use hyperspectral images as the training data and support vector regression to combine the methods. Based on the results, we confirm that their combination enhances their accuracy.

SA-MVSNet: Self-attention-based multi-view stereo network for 3D reconstruction of images with weak texture

Purpose: Benign breast disease (BBD) is viewed as a non-obligate precursor stage in the development of breast cancer (BC). Transcriptome profiling using RNA sequencing (RNA-seq) of bulk formalin-fixed paraffin-embedded (FFPE) tissue blocks has been routinely used in BC research. However, RNA-seq of heterogeneous bulk samples represents an averaged expression profile across all cell mixtures. Downstream analyses, such as biomarker discovery, are likely confounded by differences in cell proportions. There is a lack of research in optimizing strategies for cell type deconvolution especially for FFPE samples. We evaluated and optimized strategies for cell type deconvolution using existing single cell RNA-seq (scRNA-seq) and single nucleus RNA-seq (snRNA-seq) breast tissue references, and applied these to a cohort of BBD FFPE samples to assess risk of subsequent BC. Methods: 7 scRNA-seq and 3 snRNA-seq public samples from women’s breast tissue were normalized using SCTransform, then integrated using Harmony to form a complete reference set. Bioinformatics evaluations examined the performance of existing statistical and deep learning-based deconvolution methods. Computational simulations were conducted to evaluate those methods on three scenarios: 1) baseline; 2) incomplete reference set; 3) FFPE artifacts. FFPE artifacts were simulated based on a generalized additive model using 7 in-house paired FFPE-fresh frozen samples. Finally, we applied the best performing method to our BBD cohort of 62 samples. Differential expression analysis adjusting for cell proportions was performed to identify transcriptomic aberrations in BBD cases. Results: snRNA-seq was more sensitive in capturing adipocytes, while scRNA-seq was more sensitive in detecting lymphocytes in breast tissue. The median number of detectable genes per cell was higher for scRNA-seq than snRNA-seq before integration (1264 vs 558). Marker gene-based regression methods (e.g. CIBERSORT) displayed higher variance and lower bias compared to methods directly using scRNA-seq data (e.g. SCDC). Incomplete reference data led to increased bias and little effect on variance. FFPE artifacts had a higher impact on performance compared to incomplete reference data. Overall, Scaden, a deep learning-based method, was the top performing method across all simulation scenarios and was especially robust to FFPE artifacts. 59 differential expressed genes (DEGs) were identified with fold change > 1.5 and p value < 0.01. DEGs involved in BC, e.g. S100A8, WNT4 and HIST1H3H, would be left out if cell proportions were not properly adjusted. Conclusions: We constructed the first complete single cell reference data for women’s breast tissue and benchmarked existing cell type deconvolution methods. Compared to analyses not adjusting for cell proportions, we observed substantial differences in perturbed genes and pathways associated with BC risk for BBD patients. Citation Format: Yuanhang Liu, Chen Wang, Stacey J. Winham, Aditya Bhagwate, Robert A. Vierkant, Julie M. Cunningham, Bryan M. McCauley, William A. Jons, Derek C. Radisky, Amy C. Degnim. Deconvolution of cell type composition in FFPE tissue: Application to benign breast disease. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4291.

Vehicle positioning technology is essential for intelligent vehicles, supporting applications such as path planning, route guidance and autonomous parking. High-precision road topology network information is also crucial for vehicle positioning and navigation. However, approaches that generate road topology from remote sensing images are limited by viewing angle and resolution, often missing road segments and generating only road-level topology. Ground mapping vehicle systems such as simultaneous localization and mapping (SLAM) based methods struggle to directly obtain road centerlines and are rarely used directly for topology generation, while deep learning-based methods can dynamically produce lane-level topology but often lack global consistency and interpretability, suffering from redundant misconnections and high perception cost. In this paper, we present a geometry-driven framework for automatic lane-level topology generation from a monocular front-view camera, with global map alignment supported by external localization. Unlike purely deep learning-based approaches, our approach leverages existing high-precision ground vehicle mapping algorithms to extract road geometry constraints and incorporates semantic cues to ensure interpretable and globally connected topology generation, combining interpretability with strong geometric supervision. Specifically, we introduce a lane semantic information based lane enhancement strategy to fill topological gaps, a DBSCAN based three-stage pipeline to denoise and regularize lane center nodes, and a virtual-node representation to simplify complex intersections. Our approach has been experimentally validated in real-world urban areas and compared to existing mainstream deep learning-based topology generation methods. The results indicate that our method achieves higher topological accuracy and completeness within the same front-view camera field of view, demonstrating its effectiveness for lane-level topology construction.

It is one of the most challenging problems in applied mathematics to approximatively solve high-dimensional partial differential equations (PDEs). Recently, several deep learning-based approximation algorithms for attacking this problem have been proposed and tested numerically on a number of examples of high-dimensional PDEs. This has given rise to a lively field of research in which deep learning-based methods and related Monte Carlo methods are applied to the approximation of high-dimensional PDEs. In this article we offer an introduction to this field of research by revisiting selected mathematical results related to deep learning approximation methods for PDEs and reviewing the main ideas of their proofs. We also provide a short overview of the recent literature in this area of research.

Recent advances in diverse remote sensing sensors have enabled the acquisition of high-spatial-resolution imagery, offering unprecedented opportunities for cost-efficient and accurate agricultural inventory and analysis in an automated manner. Lots of studies that aim at providing an inventory of the level of each agricultural parcel have generated many methods for agricultural parcel and boundary delineation (APBD). This review article covers APBD methods for detecting and delineating agricultural parcels and systematically reviews the past and present of APBD-related research applied to remote sensing images. We conceptualize existing APBD studies as comprising three hierarchical levels: cropland identification (CI), boundary delineation (BD), and parcel segmentation (PS). With the goal to provide a clear knowledge map of existing APBD efforts, we conduct a comprehensive review of recent APBD papers to build a metadata analysis, including the algorithm, the study site, the crop type, the sensor type, the evaluation method, and so on. We categorize the methods into three classes: 1) traditional image processing methods (including pixel based, edge based, and region based), 2) traditional machine learning methods [such as random forests (RFs) and decision trees (DTs)], and 3) deep learning-based methods. With deep learning-oriented approaches contributing to a majority, we further discuss deep learning-based APBD methods, organizing them primarily by task into semantic segmentation-based and object detection-based approaches and, within each, distinguishing architectures based on convolutional neural networks (CNNs) and transformers. In addition, we discuss several APBD-related issues to further comprehend the APBD domain using remote sensing data, such as multisensor data in APBD tasks, comparisons between single-task learning and multitask learning in the APBD domain, comparisons among different algorithms and different APBD tasks, and so on. Finally, this article proposes some APBD-related applications and a few exciting prospects and potential hot topics in future APBD research. We hope this review helps researchers involved in the APBD domain to keep track of developments and trends.

Deep learning-based person re-identification methods: A survey and outlook of recent works

Conventional machine learning-based methods have been effective in assisting physicians in making accurate decisions and utilized in computer-aided diagnosis for more than 30years. Recently, deep learning-based methods, and convolutional neural networks in particular, have rapidly become preferred options in medical image analysis because of their state-of-the-art performance. However, the performances of conventional and deep learning-based methods cannot be compared reliably because of their evaluations on different datasets. Hence, we developed both conventional and deep learning-based methods for lung vessel segmentation and chest radiograph registration, and subsequently compared their performances on the same datasets. The results strongly indicated the superiority of deep learning-based methods over their conventional counterparts.

Deep learning (DL) algorithms are considered as a methodology of choice for remote-sensing image analysis over the past few years. Due to its effective applications, deep learning has also been introduced for automatic change detection and achieved great success. The present study attempts to provide a comprehensive review and a meta-analysis of the recent progress in this subfield. Specifically, we first introduce the fundamentals of deep learning methods which are frequently adopted for change detection. Secondly, we present the details of the meta-analysis conducted to examine the status of change detection DL studies. Then, we focus on deep learning-based change detection methodologies for remote sensing images by giving a general overview of the existing methods. Specifically, these deep learning-based methods were classified into three groups; fully supervised learning-based methods, fully unsupervised learning-based methods and transfer learning-based techniques. As a result of these investigations, promising new directions were identified for future research. This study will contribute in several ways to our understanding of deep learning for change detection and will provide a basis for further research. Some source codes of the methods discussed in this paper are available from: https://github.com/lazharkhelifi/deeplearning_changedetection_remotesensing_review.

<p>This thesis considers the analysis of matched filtering (MF) processing in massive multi-user multiple-input-multiple-output (MU-MIMO) wireless communication systems. The main focus is the analysis of system performance for combinations of two linear processers, analog maximum ratio combining (MRC) and digital MRC. We consider implementations of these processing techniques both at a single base-station (BS) and in distributed BS layouts. We further consider extremely low complexity distributed variants of MRC for such systems. Since MRC relies on the massive MIMO properties of favourable propagation (FP) and channel hardening, we also present a detailed analysis of FP and channel hardening. This analysis employs modern ray-based models rather than classical channel models as the models are more reliable for the large arrays and higher frequencies envisaged for future systems. The importance of MRC processing is being driven by the emergence of massive MIMO and millimetre wave as strong candidates for next generation wireless communication systems. Massive MIMO explores the spatial dimension by providing significant increases in data rate, link reliability and energy efficiency. However, with a large number of antennas co-located in a fixed physical space, correlation between the elements of antennas may have a negative impact. Distributed systems, where the total number of antennas are divided into different locations, make this problem less serious. Also, linear processing techniques, analog MRC and digital MRC, due to their simplicity and efficiency, are more practical in massive MU-MIMO systems. For these reasons we consider MRC processing in both co-located and distributed scenarios. Although distributed systems reduce the adverse impact of correlation caused by closely-spaced large antenna arrays by dividing the antennas into multiple antenna clusters, the correlation within the cluster still exists. Thus, we extend MRC analysis for massive MIMO to correlated channels. Approximations of expected per-user spectrum efficiency (SE) with correlation effects for massive MIMO systems with analog MRC and digital MRC are derived. Useful insights are given for future system deployments. A convergence analysis of the interference behaviour under different correlation models is presented. Furthermore, a distributed fully cooperative system, where all the received signals are sent to the central processor, offers attractive performance gains but at the cost of high computational complexity at the central node. Thus, we propose four low-complexity, two-stage processors, where only processed signals after local processing (first-stage) are transmitted to the global processing node (second-stage). We present analytical expressions for the expected per user SINR in an uplink distributed MU-MIMO system with two-stage beam-forming. This leads to an approximation of expected per-user SE. The analysis of both millimetre wave and massive MIMO systems requires a strong link to the physical environment and ray-based models are more practical and suitable for such systems. However, it is unclear how the key properties in conventional MIMO systems, such as FP and channel hardening, will behave in a ray-based channel model. In this thesis, remarkably simple and general results are obtained demonstrating that: a) channel hardening may or may nor occur depending on the nature of the channel models; b) FP is guaranteed for all models as long as the ray angles are continuous random variables; c) we also propose a novel system metric, denoted large system potential (LSP) as the ratio of the mean desired signal power to the total mean interference power, where both the numbers of antennas and end-users are growing to infinity at a fixed ratio. We derive simple approximations to LSP and demonstrate that LSP will not normally hold as the mean interference power usually grows logarithmically relative to the mean signal power.</p>

Cyber-physical systems (CPS) have been widely deployed in critical infrastructures and are vulnerable to various attacks. Data integrity attacks manipulate sensor measurements and cause control systems to fail, which are one of the prominent threats to CPS. Anomaly detection methods are proposed to secure CPS. However, existing anomaly detection studies usually require expert knowledge (e.g., system model-based) or are lack of interpretability (e.g., deep learning-based). In this paper, we present Deepnoise, a deep learning-based anomaly detection method for CPS with interpretability. Specifically, we utilize the sensor and process noise to detect data integrity attacks. Such noise represents the intrinsic characteristics of physical devices and the production process in CPS. One key enabler is that we use a robust deep autoencoder to automatically extract the noise from measurement data. Further, an LSTM-based detector is designed to inspect the obtained noise and detect anomalies. Data integrity attacks change noise patterns and thus are identified as the root cause of anomalies by Deepnoise. Evaluated on the SWaT testbed, Deep-noise achieves higher accuracy and recall compared with state-of-the-art model-based and deep learning-based methods. On average, when detecting direct attacks, the precision is 95.47%, the recall is 96.58%, and F is 95.98%. When detecting stealthy attacks, precision, recall, and F scores are between 96% and 99.5%.

Crop grading represents a critical component in enhancing agricultural production efficiency and optimizing supply chain management. Through systematic assessment of crop appearance, quality attributes, and dimensional characteristics, accurate determination of market compliance and evaluation of quality grades and commercial value can be achieved. This systematic review comprehensively analyzes 53 studies published between 2020 and 2025 that employed deep learning-based target detection methods for automated crop grading applications. We establish a taxonomic framework categorizing detection methodologies into three distinct paradigms: two-stage detectors (characterized by high precision but substantial computational overhead), single-stage detectors (offering optimal speed-accuracy trade-offs), and emerging Transformer-based approaches (demonstrating superior global modeling capabilities). Our analysis reveals significant algorithmic evolution trends, with the YOLO series exhibiting dominant application across more than 30 crop varieties, achieving high detection accuracy while maintaining real-time performance capabilities. We systematically identify key technical challenges: sparse feature extraction under complex agricultural environments, severe class imbalance in defect detection tasks, and deployment constraints in edge computing scenarios. We propose a comprehensive solution framework that integrates multimodal data fusion (RGB-D-NIR), lightweight architectural optimization strategies, and semi-supervised learning approaches capable of reducing annotation requirements by 60–80%. Performance benchmarking on public datasets demonstrates that target detection methods exhibit superior engineering applicability compared to segmentation approaches, with enhanced inference speed and improved robustness to environmental variations. For future research directions, we categorize priorities into four domains: intelligent annotation systems, environmental adaptability enhancement, model compression techniques, and cross-domain transfer learning. This review provides a comprehensive technical framework for deep learning-driven crop grading, offering actionable insights for researchers and practical deployment guidelines for precision agriculture applications. We anticipate this review will inspire broader adoption of deep learning target detection methodologies in agricultural domains, fostering technological advancement and innovation in this field.

Deep learning has significantly transformed personalized education by enabling intelligent adaptation to individual learning needs. This study explores deep learning-based modeling methods that enhance personalized learning experiences, optimize instructional content, and predict student progress. We examine key techniques, including recurrent neural networks (RNNs), transformers, reinforcement learning, and multimodal learning analytics, to demonstrate their roles in personalized learning path recommendations and adaptive content generation. Case studies of AI-driven tutoring systems and learning management platforms illustrate real-world applications. Additionally, we address challenges related to data privacy, algorithmic bias, and model interpretability. The paper concludes with future directions for deep learning in education, emphasizing its potential for enhancing immersive and intelligent learning environments.

Monocular human pose estimation: A survey of deep learning-based methods