High Execution Speed Research Articles

Dental segmentation via enhanced YOLOv8 and image processing techniques

By blending computer-aided medical systems with cutting-edge privacy technologies, healthcare providers can deliver more personalized, effective care while maintaining the highest data security standards and patient trust. The challenge of dental segmentation in computer vision, a task focused on accurately outlining dental structures in images, traditional methods, particularly convolution neural networks (CNNs), didn't reach high accuracy in this area due to suboptimal performance and computational inefficiency. The goal of image segmentation is to group pixels on the basis of their visual properties, such as color, texture, intensity, or spatial proximity, to identify and delineate the boundaries of distinct objects or regions within the image. In this paper, You Only Look Once (YOLOv8) algorithm is improved to segment teeth with high accuracy and high execution speed. The increase in the number of layers of YOLOv8 relied upon, as the accuracy of the algorithm segmentation depends on the number of layers used to extract features from the image (backbone) and the number of layers of the head (prediction). In addition, the size of the layers is decreased to increase the execution speed. The novelty of this paper is the proposed YOLOv8 model in addition to the Proposed Activation Function (PAF). The dataset (top view) used was taken from a dental clinic where 526 images were taken of dental and different patients. The best accuracy reached 99.561% when the enhanced YOLOv8 segmentation model was applied to the dental dataset. It can be concluded that the improved model of the YOLOv8 algorithm has increased the accuracy of dental segmentation compared to previous research because it relies on a proposed PAF that increases the difference between the features extracted from the layers of the proposed model which makes it able to distinguish between teeth and surrounding parts significantly.

Effect of middle stud on ultimate strength and behavior factor of brick masonry shear wall in cold formed steel frame

AbstractLightweight steel framing is one of the modern construction technology systems. This system is mostly used for low to mid-rise buildings. The lightweight steel framing system has many advantages, including lightness, ease of installation, high execution speed, and being more cost-efficient. Since the manufacturers of cold formed steel frames use bricks in an unprincipled way to cover these structures and because of less laboratory research in this regard, in the present research to principled use of this structure, the effect of the middle stud was evaluated on seismic behavior of brick shear wall in cold formed steel frame with brick face. For this purpose, four cold formed steel frames were made in two different configurations (without middle stud and with middle stud) using cement sand mortar, wire mesh, and brick shear walls. Based on the results, the middle stud would cause weakness in the permissible deformation of the brick walls, and in shear walls without middle stud, deformations occur along with the acquisition of resistance to larger deformations. Accordingly, the presence of the middle stud increases the average shear strength by 30%, and this increase in resistance causes a decrease in the behavior factor and ductility of the walls, which practically indicates the seismic behavior of the frames with the middle stud.

Automated License Plate Recognition for Non-Helmeted Motor Riders Using YOLO and OCR

One of the leading causes of serious injuries in motorcycle accidents is the failure to wear a helmet, pressing the need for effective measures to encourage riders to use helmets. To stop these frequent violations of business regulations, regular observation is necessary. The proposed system is for detecting traffic violations in India related to riding motorcycles without helmets. The system utilizes deep learning-based object detection using YOLO and OCR techniques to automatically detect non-helmet riders and extract license plate numbers. The proposed system aims to improve efficiency and accuracy in detecting violations by automating the process and reducing the need for manpower. The system involves three levels of object detection: person and motorcycle/moped, helmet, and license plate. OCR is used to extract the license plate number, and all techniques are subject to predefined conditions and constraints. The system is designed to operate on video input to ensure high-speed execution, and it intends to offer a complete solution for both detection of helmet and extracting the license plate number.

Evaluation of classical correlation functions from 2/3D images on CPU and GPU architectures: Introducing CorrelationFunctions.jl

Correlation functions are becoming one of the major tools for quantification of structural information that is usually represented as 2D or 3D images. In this paper we introduce ▪ open-source package developed in Julia and capable of computing all classical correlation functions based on imaging input data. Images include both binary and multi-phase representations. Our code is capable of evaluating two-point probability S2, phase cross-correlation ρij, cluster C2, lineal-path L2, surface-surface Fss, surface-void Fsv, pore-size P and chord-length p distribution functions on both CPU and GPU architectures. Where possible, we presented two types of computations: full correlation map (correlations of each point with other points on the image, that also allows obtaining ensemble averaged CF) and directional correlation functions (currently in major orthogonal and diagonal directions). Such an implementation allowed for the first time to assemble a completely free solution to evaluate correlation functions under any operating system with well documented application programming interface (API). Our package includes automatic tests against analytical solutions that are described in the paper. We measured execution times for all CPU and GPU implementations and as a rule of thumb full correlation maps on GPU are faster than other methods. However, full maps require more RAM and, thus, are limited to available RAM resources. On the other hand, directional CFs are memory efficient and can be evaluated for huge datasets – this way they are the first candidates for structural data compression of feature extraction. The package itself is available through Julia package ecosystem and on GitHub, the latter source also contains documentation and additional helpful resources such as tutorials. We believe that a single powerful computational tool such as ▪ presented in this paper will significantly facilitate the usage of correlation functions in numerous areas of structural description and research of porous materials, as well as in machine learning applications. We also present some examples as applied to ceramic, soil composite and oil-bearing rock samples based on their 3D X-ray tomography and 2D scanning electron microscope images. Finally, we conclude our paper with discussion of possible ways to further improve presented computational framework. Program summaryProgram Title: CorrelationFunctions.jlCPC Library link to program files:https://doi.org/10.17632/6gb9gfm3dw.1Developer's repository link:https://github.com/fatimp/CorrelationFunctions.jlLicensing provisions: MITProgramming language: JuliaSupplementary material: Numerous Jupiter notebooks with examples are available on the GitHub pageNature of problem: Correlation functions are invaluable universal statistical descriptors of structures used in numerous scientific fields such as astronomy, material science, rock and soil physics, hydrology and biology, to name just a handful of examples. While computational approaches are available in the literature for some functions, they are fragmented and are usually implemented in proprietary interpreted languages for CPU architecture alone.Solution method: We contribute an open source and cross-platform solution with well documented API for computation of all classical correlation functions from both 2D and 3D images on CPU and GPU architectures. The package computes correlation functions using two approaches: computation of correlation maps and computation along predefined directions. These two approaches can be thought of as an execution time - memory trade-off, but the choice may also depend on the application. The computations are based on a) fast Fourier transform with preprocessing steps such as cluster labeling or edge detection, and b) linear scan approach to evaluate correlation functions along predefined directions. Where justified, the algorithms can be executed on both CPU and GPU which results in high execution speed on modern hardware.

Implementation of Association Rules Algorithm to Identify Popular Topping Combinations in Orders

Association rule is a data mining technique to find associative rules between a combination of items. This research aims to apply association rules algorithm in identifying popular topping combinations in food orders. This application aims to help restaurant owners or food businesses understand their customers' preferences and optimize their menu offerings. Data obtained from kaggle, the association rules algorithm is applied to this dataset to identify patterns or combinations of toppings that often appear together in orders. The results of this study show toppings with chocolate as a popular item in orders. These findings can provide valuable insights for food business owners in structuring their menus and determining attractive offers for customers. This study also applied a comparison between the apriori, fp- growth and eclat algorithms, with the result that the best item transaction rule was found: a combination of dill & unicorn toppings with chocolate with 60% confidence. Overall, the application of eclat algorithm in this study provides the best performance with higher execution speed, thus providing insight into customer preferences regarding topping combinations in food orders. Despite the shortcomings of the data form from this study, it is expected to help business owners in optimizing their offerings, increasing customer satisfaction, and improving their business performance.

РОЛЬ ИСКУССТВЕННОГО ИНТЕЛЛЕКТА В БИЗНЕСЕ

Small businesses play a vital role in the fabric of any economy. The Small Business Administration recognizes their significant contributions to economic growth, job creation, innovation, and overall prosperity. These businesses, although small in scale and often privately owned, have significant value compared to larger corporations. The definition of a small business can vary from country to country and industry to industry, encompassing factors such as the number of employees, annual revenue, and assets. Although various AI technologies are used quite often in everyday life (72% of respondents use voice assistants, smart gadgets, or automatic control of home appliances), less than half of respondents (45%) have used AI in their work processes in the past year. The most popular tasks are creative ones, which usually take up a lot of employees' operational time. For example, instead of recording a meeting and then spending several hours transcribing it, speech recognition technologies are used to translate it into text. Communication with customers can be simplified with the help of smart voice assistants. In order to find analogues of your products, it is not necessary to physically go around all competitors, sometimes it is enough to upload a photo to the image recognition and analysis service. However, only 19% of entrepreneurs used the capabilities of chatbots in their work that can support a dialogue, or virtual assistants that make an action plan and write a scenario, 4% of entrepreneurs tried to work with a neural network that understands requests in Russian and English, can mix images, mix images and text, create different versions of images, and also redraw individual objects and areas in pictures. 78% of entrepreneurs who used artificial intelligence technologies in their work consider them useful for their activities. Among the main advantages, respondents noted the release of employee time for other tasks (56%), high speed of execution (50%) and money savings (35%).

Construction and verification of machine vision algorithm model based on apple leaf disease images.

Apple leaf diseases without timely control will affect fruit quality and yield, intelligent detection of apple leaf diseases was especially important. So this paper mainly focuses on apple leaf disease detection problem, proposes a machine vision algorithm model for fast apple leaf disease detection called LALNet (High-speed apple leaf network). First, an efficient sacked module for apple leaf detection, known as EALD (efficient apple leaf detection stacking module), was designed by utilizing the multi-branch structure and depth-separable modules. In the backbone network of LALNet, (High-speed apple leaf network) four layers of EALD modules were superimposed and an SE(Squeeze-and-Excitation) module was added in the last layer of the model to improve the attention of the model to important features. A structural reparameterization technique was used to combine the outputs of two layers of deeply separable convolutions in branch during the inference phase to improve the model's operational speed. The results show that in the test set, the detection accuracy of the model was 96.07%. The total precision was 95.79%, the total recall was 96.05%, the total F1 was 96.06%, the model size was 6.61 MB, and the detection speed of a single image was 6.68 ms. Therefore, the model ensures both high detection accuracy and fast execution speed, making it suitable for deployment on embedded devices. It supports precision spraying for the prevention and control of apple leaf disease.

Analysis of Methods used to Create Digital Signals Fingerprints

Digital signals are used in many vital applications, and since the volume of this data is very large, it is necessary to represent this data with a set of values and in a small size in order to increase the speed of processing the digital signal, because the applications that use this data need a high execution speed. In this paper research a detailed study of kmeans clustering and wavelet packet tree decomposition will be introduced, Using these methods to process any type of digital signals will lead to represent the digital data by a set of small number of values, this set will be used as a fingerprint to identify the digital data. Kmeans clustering method and wavelet tree decomposition will be used to create a fingerprints for speech signals, color images and gray images, the obtained results will be analyzed to give the user a bets way to create a fingerprint for any type of digital data, which is required to build any computer system, which use the digital data as an input.

A Lightweight Conditional Convolutional Neural Network for Hyperspectral Image Classification

Deep learning (dl), especially convolutional neural networks (cnns), has been proven to be an excellent feature extractor and widely applied to hyperspectral image (hsi) classification. However, dl is a computationally demanding algorithm with many parameters and a high computational burden, which seriously restricts the deployment of dl-based hsi classification algorithms on mobile and embedded systems. In this paper, we propose an extremely lightweight conditional three-dimensional (3D) hsi with a double-branch structure to solve these problems. Specifically, we introduce a lightweight conditional 3D convolution to replace the conventional 3D convolution to reduce the computational and memory cost of the network and achieve flexible hsi feature extraction. Then, based on lightweight conditional 3D convolution, we build two parallel paths to independently exploit and optimize the diverse spatial and spectral features. Furthermore, to precisely locate the key information, which is conducive to classification, a lightweight attention mechanism is carefully designed to refine extracted spatial and spectral features, and improve the classification accuracy with less computation and memory costs. Experiments on three public hsi data sets show that the proposed model can effectively reduce the cost of computation and memory, achieve high execution speed, and better classification performance compared with several recent dl-based models.

Programmable photonic integrated meshes for modular generation of optical entanglement links

Large-scale generation of quantum entanglement between individually controllable qubits is at the core of quantum computing, communications, and sensing. Modular architectures of remotely-connected quantum technologies have been proposed for a variety of physical qubits, with demonstrations reported in atomic and all-photonic systems. However, an open challenge in these architectures lies in constructing high-speed and high-fidelity reconfigurable photonic networks for optically-heralded entanglement among target qubits. Here we introduce a programmable photonic integrated circuit (PIC), realized in a piezo-actuated silicon nitride (SiN)-in-oxide CMOS-compatible process, that implements an N × N Mach–Zehnder mesh (MZM) capable of high-speed execution of linear optical transformations. The visible-spectrum photonic integrated mesh is programmed to generate optical connectivity on up to N = 8 inputs for a range of optically-heralded entanglement protocols. In particular, we experimentally demonstrated optical connections between 16 independent pairwise mode couplings through the MZM, with optical transformation fidelities averaging 0.991 ± 0.0063. The PIC’s reconfigurable optical connectivity suffices for the production of 8-qubit resource states as building blocks of larger topological cluster states for quantum computing. Our programmable PIC platform enables the fast and scalable optical switching technology necessary for network-based quantum information processors.

An effective authentication scheme for a secured IRIS recognition system based on a novel encoding technique

Iris recognition systems are well known for their accuracy than any other biometric system and very popular due to the permanent nature of the iris image. Moreover, it is easy to acquire the iris image even from a very long distance without causing much degradation to the image. In terms of the security perspective, very few biometric systems comprehensively address the vulnerabilities occurring as a result of security attacks. Among the security vulnerabilities, the possibility of compromise of Iris Code during the exchange of Iris information between client and server is a serious threat. In this work, an innovative model which generates a One Time Iris Code (OTIC) for every user authentication is proposed with the help of a novel encoding scheme. In the proposed model, the Iris Code generated distorted using pseudo-random number generator, timestamp, and rotation parameter and it will be different for each authentication session. The proposed technique will also reduce the possibility of compromising the original Iris Code. This will eventually enhance the security of the Iris recognition system. Moreover, the proposed method has less time complexity and high execution speed than the conventional Iris recognition systems. The accuracy of this model is comparable with the existing systems. Various performance analysis methods such as histogram processing for showing the pixel intensity of iris images, and calculation of EER based on entropy measurement for evaluating the severity of attacks using Base Score calculation have been implemented in this work.

Quantum Dots‐Functionalized Encryption Camera for Text Image Security Protection

With the wide utilization of mobile devices with a camera as online communication tools, there is a great need to develop image encryption technology for protecting confidential and sensitive information. Herein, quantum dots‐functionalized encryption (QDE) camera is proposed by integrating the fluorescence emission of quantum dots (QDs) with charge coupled device or complementary metal oxide semiconductor detector arrays during the imaging process. The variation of the photoluminescence emission from QDs controlled by UV365 nm illumination provides a convenient way to scramble the spatial distribution of RGB values in the captured image, which expands the information entropy of the captured image to realize in situ image encryption. The encrypted text images can be decrypted by a convolutional neural network‐based algorithm, resulting in high‐quality decrypted images with peak signal‐to‐noise ratio up to 45 dB. Compared with the previous reported image encryption techniques of chaotic cryptosystem, fluorescent material‐based encryption, and optical encryption, the proposed QDE camera shows advantages of high execution speed and easy miniaturization. It provides a promising in situ image encryption technique to guarantee information security in digitization process.

Road Recognition Based on Vehicle Vibration Signal and Comfortable Speed Strategy Formulation Using ISA Algorithm.

When a vehicle is being driven, it is excited by the road roughness and generates its own vibration. In order to improve the vehicle’s riding comfort and the physical–mental health of passengers in the vehicle, this paper proposes a formulation method for a comfortable speed strategy and the technical route of its application. According to international standard ISO 2631-1, the relationship between the weighted root-mean-square acceleration value and comfortable vehicle speed is analyzed. The simulation test platform of the road roughness signal and vehicle vibration signal is built by using the filtering white noise method and the second Lagrange equation through Matlab/Simulink. Combined with the simulation platform, this paper extracts seven characteristics with statistical properties from the time-domain signal and obtains 500 sample data. Random forest (RF), extreme learning machine (ELM), and radial basis function neural network (RBF-NN) are applied to identify roads. Two comfortable speed strategy formulation methods based on the improved simulated annealing (ISA) algorithm are proposed and compared according to the solution effect of each grade of comfortable speed. The results show that the simulated signals of each grade road roughness are accurate. Road recognition can be effectively carried out using the statistical characteristics of vehicle vibration acceleration signals. ELM has high recognition accuracy and fast execution speed. The ISA-II algorithm has a low solution error of comfortable speed and a low computation time. The comfortable speed of the research vehicle on different road grades showed a great difference.

GenomicSimulation: fast R functions for stochastic simulation of breeding programs.

Simulation tools are key to designing and optimizing breeding programs that are multiyear, high-effort endeavors. Tools that operate on real genotypes and integrate easily with other analysis software can guide users toward crossing decisions that best balance genetic gains and genetic diversity required to maintain gains in the future. Here, we present genomicSimulation, a fast and flexible tool for the stochastic simulation of crossing and selection based on real genotypes. It is fully written in C for high execution speeds, has minimal dependencies, and is available as an R package for the integration with R’s broad range of analysis and visualization tools. Comparisons of a simulated recreation of a breeding program to a real data set demonstrate the simulated offspring from the tool correctly show key population features, such as genomic relationships and approximate linkage disequilibrium patterns. Both versions of genomicSimulation are freely available on GitHub: The R package version at https://github.com/vllrs/genomicSimulation/ and the C library version at https://github.com/vllrs/genomicSimulationC/.

Efficient and lightweight detection of PPG onset and systolic peaks using implementable time-domain strategies

A majority of clinically reliable Photoplethysmogram (PPG) signal features necessitate exact location of PPG onset and systolic peaks. So far, state-of-the-art researches in the field of PPG peak detection is lagging due to the use of amplitude thresholding, complicated techniques or pathophysiologically limited datasets. In this paper, a robust algorithm is developed that primarily uses a novel combination of PPG-derivative, trigonometric area of the curve (TAOC), slope reversal and time-domain based lightweight methodologies to uncover the precise location of PPG onset and systolic peaks. Rigorous evaluation of the algorithm is conducted over 2,16,374 beats, collected from standard databases together with the PPG signal acquired from healthy and cardiac subjects. Superiority of the algorithm is presented by high average accuracy, sensitivity, predictivity and error as 99.69%, 99.77%, 99.92% and 0.31%. The adopted lightweight methodologies and high-speed execution over a wide pathophysiological variety presents immense promises for implementation in standard healthcare applications.

Short-wavelength reverberant wave systems for physical realization of reservoir computing

Machine learning (ML) has found widespread application over a broad range of important tasks. To enhance ML performance, researchers have investigated computational architectures whose physical implementations promise compactness, high-speed execution, physical robustness, and low energy cost. Here, we experimentally demonstrate an approach that uses the high sensitivity of reverberant short wavelength waves for physical realization and enhancement of computational power of a type of ML known as reservoir computing (RC). The potential computation power of RC systems increases with their effective size. We here exploit the intrinsic property of short wavelength reverberant wave sensitivity to perturbations to expand the effective size of the RC system by means of spatial and spectral perturbations. Working in the microwave regime, this scheme is tested experimentally on different ML tasks. Our results indicate the general applicability of reverberant wave-based implementations of RC and of our effective reservoir size expansion technique

Model predictive control for optimal power flow in grid-connected unbalanced microgrids

This paper presents a model predictive control (MPC) designed to solve the optimal power flow (OPF) in grid-connected unbalanced microgrids. The controller considers the modeling of distributed renewable generators, storage units, unbalanced loads, voltage regulators, tap transformers and capacitor banks with tap controls. Active and reactive power flows are optimized within the same optimization problem, thus allowing renewable generators to produce reactive power. A predictive OPF is modeled as a mixed integer non-linear programming problem and a convenient linearization is proposed, so that, the OPF can be implemented through commercial solvers. Testing is performed in a 24 h simulation in modified versions of the IEEE 13 and IEEE 123 bus test feeders, and the results are compared with recent controllers. The controller is capable of finding optimum operation points according to predictable changes updated online, while still maintaining a high execution speed required for real-time applications. This tool performs an accurate OPF with reduced computational effort, with the aim of saving resources in control systems sizing and operation.

Multi-shuttle AS/RS dimensions optimization using a genetic algorithm—case of the multi-aisle configuration

Over the past few years, automated storage and retrieval systems (AS/RSs) have been increasingly improving. It is worth mentioning that multi-shuttle storage/retrieval (S/R) machines were gradually introduced to the market some years ago. These machines, which possess a high speed of execution, are able to transport several pallets at a time during the same trip, as opposed to single-shuttle S/R machines, which can carry only one pallet at a time. It should be noted that the installation of this type of system requires a significant financial investment, and therefore, it is highly recommended that this system be well studied and designed prior to its installation. It is widely acknowledged that one of the most important objectives while designing an AS/RS is to achieve the shortest time for one single cycle. The present work aims at designing an AS/RS with optimal dimensions for the purpose of minimizing the time in a multi-cycle implementation. To do this, it was decided to consider a multi-aisle automated storage/ retrieval system (AS/RS) with a multi-shuttle S/R machine. In addition, a genetic algorithm (GA) was used for the optimization of the system.

A Proposed Method for Achieving the Confidentiality of Arabic Texts

Due to the rapid development of distributed computing systems, the use of local networks, and the huge expansion of the Internet, the process of maintaining the confidentiality of information becomes important and necessary. Encryption is one of the most important ways to keep information confidential and to prevent unauthorized people from disclosing this information. Many cryptographic algorithms have appeared in different ways and methods. But most of these algorithms have been devoted for encrypting texts in the English language. Because of the increasing important and sensitive information exchange in the Arabic language by users on the Internet, an urgent need has appeared to build dedicated encryption systems for the Arabic language. The aim of this research is to provide a proposed encryption method that uses the idea of combining the methods of transposition and substitution cipher for the purpose of Arabic texts confidentiality. This research depends on using diacritics in the Arabic language to perform the process of substitution cipher. The efficiency of the proposed method has been tested on different texts; the method provides high speed execution, the length of the encrypted texts is almost close to the clear text length, and the confidentiality of the resulting encrypted texts is high.

Row anchor selection classification method for early-stage crop row-following

The field navigation tasks of agricultural robots are the basis for implementing field management projects in the early stages of seedlings guided by precision agriculture policies. High accuracy and execution speed of navigation tasks are needed for the escalating field machine implementations in a significantly unstructured and irregular field environment, contributing to the difficulty and urgency of achieving a field navigation solution with high navigation accuracy while maintaining proper execution speed. In our study, the row anchor selection classification method (RASCM) is presented to track early-stage crop rows, and the early corn seedling dataset (E-dataset) and LRC-dataset are developed for the measurement of the experimental results. RASCM is divided into three parts, including a row anchor selection method (RasM), a lightweight deep convolutional neural network (DCNN) network GhostNet, and a combined loss function (C-loss). It is suggested that RASCM has achieved good results in terms of calculation speed and accuracy by comparing the end-to-end regression operation based on the steering wheel angle (E2Er) and crop row pixel segmentation (SeG) approaches with RASCM on the LRC-dataset, which can be effectively applied in actual projects. Finally, a class activation map (CAM) is used to visualize the internal working mechanism of GhostNet, and the demonstrated effectiveness of GhostNet is not limited to the RasM labeling area. In contrast, broad global features are learned, and the previous noise is highly mitigated. Strong generalization performance for different environments is displayed by using RACSM, contributing to the effectiveness of early-stage crop row-following. The main contribution of our paper is the proposal of RASCM, which is a novel method in the field of early-stage crop row-following and can be used as a basic reference for subsequent work in this research field.