Implementation Of Algorithm Research Articles

LiDAR Loop Closure Detection using Semantic Graphs with Graph Attention Networks

In this paper, we propose a novel loop closure detection algorithm that uses graph attention neural networks to encode semantic graphs to perform place recognition and then use semantic registration to estimate the 6 DoF relative pose constraint. Our place recognition algorithm has two key modules, namely, a semantic graph encoder module and a graph comparison module. The semantic graph encoder employs graph attention networks to efficiently encode spatial, semantic and geometric information from the semantic graph of the input point cloud. We then use self-attention mechanism in both node-embedding and graph-embedding steps to create distinctive graph vectors. The graph vectors of the current scan and a keyframe scan are then compared in the graph comparison module to identify a possible loop closure. Specifically, employing the difference of the two graph vectors showed a significant improvement in performance, as shown in ablation studies. Lastly, we implemented a semantic registration algorithm that takes in loop closure candidate scans and estimates the relative 6 DoF pose constraint for the LiDAR SLAM system. Extensive evaluation on public datasets shows that our model is more accurate and robust, achieving 13% improvement in maximum F1 score on the SemanticKITTI dataset, when compared to the baseline semantic graph algorithm. For the benefit of the community, we open-source the complete implementation of our proposed algorithm and custom implementation of semantic registration at https://github.com/crepuscularlight/SemanticLoopClosure.

A full-scale simulator for research of navigation and filtering algorithms of a strapdown inertial navigation system using the Matlab-Simulink environment

Due to the increasing complexity of aircraft navigation equipment and the growing demands placed on them, there is a need to study and improve existing navigation and filtering algorithms by solving problems of developing full-scale research simulators. The article presents the results of work in the field of creating a full-scale simulator for research of navigation and filtering algorithms for a strapdown inertial navigation system (SINS) comprising: primary navigation data sensors made using microelectromechanical system technology (MEMS), servos and a navigation platform with two-degrees-of-freedom in roll and pitch. The article presents the features of the design, hardware and algorithmic implementation of the test rig taking into account the prospects for its development in terms of using the number of degrees of freedom of the platform (pitch, roll and yaw channels). The implemented principle of integrating the Simulink model of the control object is described. The control object consists of a controller based on the Arduino platform, a GPS sensor, a GY-91 sensor with an inertial measurement unit consisting of three orthogonally located: angular velocity meter, accelerometer and the single-channel barometer based on the MP280 MEMS. An algorithm for positional (manual) control of the navigation platform by pitch and roll angles using two servos, through a control stick and a virtual COM port is implemented. A diagram illustrating the logic of interaction of the structural elements of the simulator, a part of the software implementation of the complementary filter used, as well as the function of its calculation and simulation links of the Simulink model are presented. The information exchange between the PC and the Arduino microcontroller is considered. A conclusion was made about the feasibility of creating and using the developed simulator to justify the use of a particular navigation and filtering algorithm for a specific type of aircraft.

Development and routine implementation of deep learning algorithm for automatic brain metastases segmentation on MRI for RANO-BM criteria follow-up.

The RANO-BM criteria, which employ a one-dimensional measurement of the largest diameter, are imperfect due to the fact that the lesion volume is neither isotropic nor homogeneous. Furthermore, this approach is inherently time-consuming. Consequently, in clinical practice, monitoring patients in clinical trials in compliance with the RANO-BM criteria is rarely achieved. The objective of this study was to develop and validate an AI solution capable of delineating brain metastases (BM) on MRI to easily obtain, using an in-house solution, RANO-BM criteria as well as BM volume in a routine clinical setting. A total of 27456 post-Gadolinium-T1 MRI from 132 patients with BM were employed in this study. A deep learning (DL) model was constructed using the PyTorch and PyTorch Lightning frameworks, and the UNETR transfer learning method was employed to segment BM from MRI. A visual analysis of the AI model results demonstrates confident delineation of the BM lesions. The model shows 100% accuracy in predicting RANO-BM criteria in comparison to that of an expert medical doctor. There was a high degree of overlap between the AI and the doctor's segmentation, with a mean DICE score of 0.77. The diameter and volume of the BM lesions were found to be concordant between the AI and the reference segmentation. The user interface developed in this study can readily provide RANO-BM criteria following AI BM segmentation. The in-house deep learning solution is accessible to everyone without expertise in AI and offers effective BM segmentation and substantial time savings.

An efficient PG-INLA algorithm for the Bayesian inference of logistic item response models

In this paper, we propose a Bayesian PG-INLA algorithm which is tailored to both one-dimensional and multidimensional 2-PL IRT models. The proposed PG-INLA algorithm utilizes a computationally efficient data augmentation strategy via the Pólya-Gamma variables, which can avoid low computational efficiency of traditioanl Bayesian MCMC algorithms for IRT models with a logistic link function. Meanwhile, combined with the advanced and fast INLA algorithm, the PG-INLA algorithm is both accurate and computationally efficient. We provide details on the derivation of posterior and conditional distributions of IRT models, the method of introducing the Pólya-Gamma variable into Gibbs sampling, and the implementation of the PG-INLA algorithm for both one-dimensional and multidimensional cases. Through simulation studies and an application to the data analysis of the IPIP-NEO personality inventory, we assess the performance of the PG-INLA algorithm. Extensions of the proposed PG-INLA algorithm to other IRT models are also discussed.

Information model of the general methodology as a system for building any activity

The research offers an information model of the general methodology as a system for building any activity. Modeling has long been an effective tool in building various processes. And since almost all processes have an information basis, information models are widely used, which are often brought to algorithmic and computer implementation. Computer technologies allow not only to quickly calculate process parameters and expected results, but also to visually display the course of the process itself with varying degrees of visualization. The greatest difficulty is modeling social processes, as they are directly related to the human factor, mostly due to the unpredictability of behavior and decisions made, often contrary to conventional logic. This is due to the lack of study of the spiritual sphere of man, his consciousness. The authors rely on the following definitions: methodology is a system of knowledge about the structure, effective organization, principles, methods and means of conscious activity; knowledge is a feature of a certain amount of information that determines its status and separates it from all other information by the criterion of its ability to solve the task; beliefs - generalizations that a person has made about cause-and-effect relationships, values, about himself/herself, other people, actions, identity, etc., what he/she considers “true” at the moment, including the conscious needs of the individual that motivate him/her to act in accordance with the accepted value orientations. The methodology as a system of activity involves the following components: knowledge - elements of the system (worldview, values, goals, principles, methods, ways and forms); human potential (according to their personal qualities and abilities, people form an organizational and functional structure and the corresponding hierarchy); interaction organized in a certain structure (management, technology, provision, execution, control); results obtained (indicators, criteria, evaluation) and feedback. Thus, the model covers all the components of any activity and allows determining the optimal methodology for building the entire process. The model is based on the quality of knowledge, which is an element of the methodological system and allows to build any activity with a high degree of adequacy and efficiency.

FPGA implementation of the HL-LHC CMS Drift Tubes Level-1 Trigger algorithm

The High-Luminosity Large Hadron Collider (HL-LHC) has motivated a generalized upgrade in electronic systems across all experiments. In the new electronics architecture for the CMS (Compact Muon Solenoid) Drift Tubes detector, the trigger generation moves from on-detector ASICs to the back-end, to be carried out by top-range FPGAs. The new algorithm aims to deliver full-resolution, offline-grade performance in the reconstruction of muon segments. To achieve this objective, meeting the latency and data rate requirements, a high-speed, highly-pipelined FPGA design with several optimizations has been developed. This work describes the architecture and performance of this algorithm, as well as the challenges encountered during implementation and the solutions adopted.

VLSI Implementation of AES Block Cipher Based Data Hiding in Image Processing

A common symmetric key block cypher for protecting electronic data is called the Advanced Encryption Standard (AES). The implementation of the AES algorithm using Very-Large-Scale Integration (VLSI) is examined in this study. Cryptographic algorithms can be realised in hardware thanks to VLSI, which has advantages over software implementations in terms of increased processing speed and security. In this work, an AES block cypher core designed and implemented with VLSI techniques is described. We investigate various architectural designs to maximise performance indicators such as area, power consumption, and throughput. The trade-offs between different implementation options and design concerns for important components such as S-boxes will be covered. To illustrate the performance and usefulness of the implemented AES core, simulation results will be given. This paper examines the VLSI implementation of the Advanced Encryption Standard (AES), with a focus on image processing applications, even though it is essential for general data security. Effective on-chip encryption and decryption can be included into image processing systems by implementing the AES algorithm into specialised hardware circuits. Benefits of this VLSI design include enhanced performance over software-based solutions on resource-constrained image processing devices, the ability to encrypt images in real-time for secure transmission or storage, and the potential for reduced power usage for battery-powered applications.

Implementation of PSK modulation and demodulation algorithms

Implementation of PSK modulation and demodulation algorithms

Research and implementation of inverse time overcurrent protection algorithm for engine interface unit

Research and implementation of inverse time overcurrent protection algorithm for engine interface unit

Optimized Algorithms for FPGA Implementation of PDCCH Chain for 5G-NR Base Station

Optimized Algorithms for FPGA Implementation of PDCCH Chain for 5G-NR Base Station

Efficient algorithm design and implementation for heterogeneity detection in glioma stem cells

Efficient algorithm design and implementation for heterogeneity detection in glioma stem cells

From intensive care monitors to cloud environments: a structured data pipeline for advanced clinical decision support.

Clinical decision-making is increasingly shifting towards data-driven approaches and requires large databases to develop state-of-the-art algorithms for diagnosing, detecting and predicting diseases. The intensive care unit (ICU), a data-rich setting, faces challenges with high-frequency, unstructured monitor data. Here, we showcase a successful example of a data pipeline to efficiently move patient data to the cloud environment for structured storage. This supports individual patient analysis, enables largescale retrospective research, and the development of data-driven algorithms. Since June 2021, ICU data of the Amsterdam UMC have been collected and stored in a third-party cloud environment which is hosted on large virtual servers. The feasibility of the pipeline will be demonstrated with the available data through research and clinical use cases. Furthermore, privacy, safety, data quality, and environmental impact are carefully considered in the cloud storage transition. Over two years, data from over 9000 patients have been stored in the cloud. The availability, agility, computational power, high uptime, and streaming data pipelines allow for large retrospective analyses as well as the opportunity to implement real-time prediction of critical events with machine learning algorithms. Critical events can be accessed by applying keyword search in the natural language data, annotated by the treating team. Besides, the cloud environment offers storage of institutional data enabling evaluation of healthcare. The combined data and features of cloud environments offer support for predictive algorithm development and implementation, healthcare evaluation, and improved individual patient care. University of Amsterdam Research Priority Agenda Program AI for Heath Decision-Making.

Flexible inference in heterogeneous and attributed multilayer networks.

Networked datasets can be enriched by different types of information about individual nodes or edges. However, most existing methods for analyzing such datasets struggle to handle the complexity of heterogeneous data, often requiring substantial model-specific analysis. In this article, we develop a probabilistic generative model to perform inference in multilayer networks with arbitrary types of information. Our approach employs a Bayesian framework combined with the Laplace matching technique to ease interpretation of inferred parameters. Furthermore, the algorithmic implementation relies on automatic differentiation, avoiding the need for explicit derivations. This makes our model scalable and flexible to adapt to any combination of input data. We demonstrate the effectiveness of our method in detecting overlapping community structures and performing various prediction tasks on heterogeneous multilayer data, where nodes and edges have different types of attributes. Additionally, we showcase its ability to unveil a variety of patterns in a social support network among villagers in rural India by effectively utilizing all input information in a meaningful way.

Clairvoyant performance bounds for adaptive beamforming in pulse-echo imaging.

In adaptive beamforming, the array signal processing adjusts its sensor delays and weights based on the incoming data. In conventional beamforming, these parameters are instead given from a predefined model. Adaptive beamformers can improve measurement precision by dynamically rejecting spatial interference. While an established theory is available on the behavior of adaptive beamformers in textbook scenarios, their expected performance on realistic pulse-echo imaging scenes is still mostly uncharted. Imaging performance can be evaluated by individual pixel precision and aggregated metrics such as resolution and contrast. The achievable gain is strongly related to the sparsity of the scene and the availability of data to appropriately estimate the spatial covariance matrix. In pulse-echo measurements, the nonstationary interference poses a special problem for adaptive beamforming, which is a current research question of academic and industrial interest. The current work establishes a performance bound for adaptive beamforming in simulated realistic pulse-echo scenarios. This is derived and numerically implemented as the clairvoyant minimum variance distortionless response beamformer. The proposed framework allows for an a priori assessment of the applicability of adaptive beamforming, for a given scenario. The performance of the implemented algorithms can be directly compared with the theoretical limit in a simulated environment.

Implementation of machine learning algorithms to identify stress and anxiety levels in the personnel of a northern public mobility company

This scientific article analyzes the use of machine learning techniques to identify levels of stress and anxiety among workers of the Empresa Pública de Movilidad del Norte. The problem lies in the high levels of stress and anxiety present in the staff. The company, committed to the prevention of occupational risks and the physical and mental well-being of its employees, implemented a control system based on an analytical-synthetic approach. These techniques processed data, identified patterns and improved their accuracy with each analysis. Through interviews with the manager, head of Human Resources and occupational physician, along with employee surveys, stress control and quality of work life were improved

Competing all-pairs shortest paths algorithms for sparse / dense graphs: implementation and comparison

In this paper we consider two families of competing algorithms for finding the shortest paths between all pairs of vertices (APSP) in directed weighted large graphs with different edge densities: Dijkstra and Floyd-Warshall. For comparison, we have taken Dijkstra's algorithm with dynamically varying binary heap, which solves the APSP prob-lem purely in parallel by repeatedly executing on all vertices of the graph considered as source vertices, and we have taken blocked Floyd-Warshall algorithm, which is also well-parallelizable. It is known that in terms of computational complexity, the first algorithm is preferable on sparse graphs and the second algorithm is preferable on dense graphs. At the same time, it is not clear what are the ranges of graph densities at which the first algorithm will consume less CPU time than the second algorithm. This paper describes multithreaded implementations of parallel algorithms on multicore processors that make different usage of synchronization primitives such as mutex, conditional variable, lock-ing, and atomic operation. By conducting computational experiments on an 8-core Intel(R) Core(TM) i7-10700 CPU @ 2.90GHz, we found that each algorithm has a preferred graph density. In the case of multi-threaded parallel imple-mentation, the blocked Floyd-Warshall algorithm has lower running time than Dijkstra's algorithm if the graph densi-ty is greater than 0.5. Otherwise, Dijkstra's algorithm runs faster. In the case of single-threaded implementation, the split point is 0.43.

Generation, evaluation and comparison of DNA-based (pseudo) random sequences

The object of this study is the methods and algorithms for computing, evaluating, and comparing DNA-based pseudorandom sequences (PRSs) and random sequences (RSs). This paper addresses the task of extracting (P)RSs with the required stochastic and statistical properties from a DNA noise source, experimentally validating these properties in compliance with the requirements of current international standards, as well as evaluating and comparing the sequences obtained for different DNA samples. The results are the developed algorithms for generating DNA-based (P)RSs, improved algorithms for their comparison, and proposals for the process of evaluating their properties. For the output of implemented algorithms, more than 96 % of the bit streams of sequences successfully pass all statistical tests, and the entropy per bit for RSs is close to 1. A special feature of the developed generation algorithms is the use of validated conditioning component – block cipher in CTR mode – which explains the possibility of obtaining unique random data with required properties. The peculiarity of the proposed evaluation algorithm is the complexity of the tests and checks used: complete assessment of statistical properties and entropy values. Special features of the improved comparison algorithms are resource saving and the ability to evaluate much larger data sets. This is due to the use of structures and data types that are better in terms of memory usage and flexible cryptographic primitives with different modes. The results could be practically applied to constructs where randomness is required: for computing the keys and system-wide parameters, when performing transformations as part of hash functions, while obtaining sequences of an arbitrary alphabet, in zero-knowledge protocols, etc.

Очистка полуструктурированных и неструктурированных данных дистанционного зондирования Земли

Abstract. The problem of processing semi-structured and unstructured Earth remote sensing (ERS) data obtained by various methods, including satellites and unmanned aerial vehicles, is touched upon. The purpose of the article is to study and implement algorithms for effective purification and preprocessing of semi-structured and unstructured Earth remote sensing data. The importance of cleaning these data from noise, artifacts and errors is emphasized in order to increase their accuracy and significance in applied scientific research and practical application. Key data preprocessing techniques are considered, including noise removal, distortion correction, classification, segmentation and standardization of data, reinforcing theoretical positions with practical examples in Python using libraries such as GDAL, OpenCV and scikit-image. Examples of detection of aerospace and transport objects using machine learning and deep learning are given, emphasizing the importance of accuracy, completeness and F1-Score metrics in assessing the quality of data purification. The practical significance of the study lies in evaluating the effectiveness of data purification methods used to restore images during remote sensing of the Earth.

Modern levels for structural strength assessment and the algorithm for implementation of the risk analysis methodology in the operation of welded metal structures

Modern levels for structural strength assessment and the algorithm for implementation of the risk analysis methodology in the operation of welded metal structures

Radiation Diagnosis of the Right Kidney Calyx Diverticulum Complicated by Urolithiasis (clinical case report)

Renal calyx diverticulum is a cystic cavity located within the kidney, lined with transitional epithelium from the inside and communicating with the calyx or with the renal pelvis through a narrow neck. Renal calyx diverticulum is a congenital pathology of the kidney, diagnosed mainly by ultrasound, computer or magnetic resonance imaging of the abdominal organs, kidneys and the retroperitoneal space. The purpose of this work was to demonstrate a case from the clinical practice of primary radiation diagnosis and differentiation of diverticulum of the right kidney calyx complicated by urolithiasis in a young woman of reproductive age. Subsequently, a successful surgical excision of the kidney calyx diverticulum was performed. The presented clinical example confirms that the use of modern methods of radiation diagnostics (ultrasound, native and intravenous contrast computed tomography) and implementation of a diagnostic algorithm for suspected diverticula of the renal calyx against the background of urolithiasis made it possible to avoid errors and contributed to the successful recuperation of the patient.