Box Size Research Articles

Quasi-2D and 3D phase-field simulation studies of polar vortex domain structures for high-density domain engineering.

Phase-field simulations were performed on the polar vortex array structure appearing in the SrTiO3/PbTiO3/SrTiO3 heterostructure to investigate the Kittel law relationship (w ∝ d1/2) between the vortex array period (w) and the PbTiO3 ferroelectric layer thickness (d). Quasi-two-dimensional simulation results show that the equilibrium period can be determined from the relationship between the size of the simulation box and the total free energy density and that the obtained period obeys the Kittel law. However, three-dimensional simulation results show that the polar vortex arrays are not homogeneous in the direction perpendicular to the vortex plane. In the pure vortex state, no obvious correlation exists between the simulation size and energy because of the dislocations caused by the vortex period mismatch. However, in the mixed state with ferroelectric a1/a2, the vortex domains exhibit equilibrium periods that satisfy the Kittel law relationship because of the confinement effect caused by phase separation. Further study of the confined vortex domains in these mixed states and the vortex domains of finite size due to mismatched dislocations are expected to enable domain engineering of higher domain densities.

Instability and warping in vertically oscillating accretion discs

Abstract Many accretion discs have been found to be distorted: either warped due a misalignment in the system, or non-circular as a result of orbital eccentricity or tidal deformation by a binary companion. Warped, eccentric, and tidally distorted discs are not in vertical hydrostatic equilibrium, and thus exhibit vertical oscillations in the direction perpendicular to the disc, a phenomenon that is absent in circular and flat discs. In extreme cases, this vertical motion is manifested as a vertical ‘bouncing’ of the gas, potentially leading to shocks and heating, as observed in recent global numerical simulations. In this paper we isolate the mechanics of vertical disc oscillations by means of quasi-2D and fully 3D hydrodynamic local (shearing-box) models. To determine the numerical and physical dissipation mechanisms at work during an oscillation we start by investigating unforced oscillations, examining the effect of initial oscillation amplitude, as well as resolution, boundary conditions, and vertical box size on the dissipation and energetics of the oscillations. We then drive the oscillations by introducing a time-dependent gravitational potential. A key result is that even a purely vertically oscillating disc is (parametrically) unstable to developing inertial waves, as we confirm through a linear stability analysis. The most important of these has the character of a bending wave, whose radial wavelength depends on the frequency of the vertical oscillation. The nonlinear phase of the instability exhibits shocks, which dampen the oscillations, although energy can also flow from the bending wave back to the vertical oscillation.

Laboratory Study of Interface Friction between Sand and Steel Surfaces

Abstract In this study, results have been reported from a series of laboratory tests to determine the interface friction angle of steel surfaces with sand through modified direct shear testing. For completion, steel surfaces with different surface geometry to replicate the roughness have been slid against various sands compacted at different relative densities under both dry and wet conditions. The standard square shape direct shear box has been modified into circular and rectangular boxes to sufficiently enhance the contact area and to subsequently quantify its effect on interface friction angle. A mild steel interface of smooth and rough texture with a groove depth of 0.2 mm and tipping at an angle of 90° has been used in this study. The analysis of results revealed that the interface friction angles obtained from the standard direct shear box agree closely with those obtained from both modified rectangular and circular shear boxes. However, the ratio of the interface friction angle from the modified apparatus and soil’s angle of internal friction significantly depends upon the roughness of the surface in contact with the sand, its median particle size, and is independent of the effects of box size, soil’s relative density, and dry or wet condition. A comparison between existing well-accepted guidelines and current observations has been presented for use by the practicing engineers dealing with soil-structure interaction.

Observation of Two-Dimensional Dam Break Flow and a Gaseous Phase of Solitons in a Photon Fluid.

We report the observation of a two-dimensional (2D) dam break flow of a photon fluid in a nonlinear optical crystal. By precisely shaping the amplitude and phase of the input wave, we investigate the transition from one-dimensional (1D) to 2D nonlinear dynamics. We observe wave breaking in both transverse spatial dimensions with characteristic timescales determined by the aspect ratio of the input box-shaped field. The interaction of dispersive shock waves propagating in orthogonal directions gives rise to a 2D ensemble of solitons. Depending on the box size, we report the evidence of a dynamic phase characterized by a constant number of solitons, resembling a 1D soliton gas in integrable systems. We measure the statistical features of this gaslike phase. Our findings pave the way to the investigation of collective solitonic phenomena in two dimensions, demonstrating that the loss of integrability does not disrupt the dominant phenomenology.

Fully automated detection of anomalous aortic origin of the coronary arteries in coronary CT images: a novel artificial intelligence-based screening tool

Abstract Background Anomalous Aortic Origin of the Coronary Artery (AAOCA) is a rare congenital disease that can be easily overlooked during contrast-enhanced Coronary Computed Tomography Angiography (CCTA) image analysis. Objective In this study, we aimed to develop a novel artificial intelligence-based screening tool to automate the detection of AAOCA in CCTA images. Method We enrolled 127 patients from the coronary artery anomaly registry from two centers with AAOCA diagnosis based on CCTA as well as 413 normal cases. A deep learning model was developed to automatically segment the aorta and the left ventricle (LV), after which the images were automatically cropped to the aortic root and LV intersection with a box size of 6 cm3. The images were then clipped to Hounsfield Units ranging from -400 to 750 and normalized to -1 to1. These cropped regions were fed into different residual convolutional neural networks (ResNet, with different depths 10, 18, 34 and 50) to detect AAOCA. Model development (parameter and hyperparameters optimization) was performed on the training and validation sets (70/10%), and finally, it was tested on the 20% untouched test set. Various classification metrics were reported for the detection of AAOCA. Results ResNet-10 achieved an AUC of 0.91 (sensitivity: 0.89, specificity: 0.86), while ResNet-18 showed improved performance with an AUC of 0.95 (sensitivity: 0.89, specificity: 0.95). ResNet-34 achieved an AUC of 0.94 (sensitivity: 0.83, specificity: 0.96), and ResNet-50 excelled with the highest AUC of 0.96 (sensitivity: 0.89, specificity: 0.95), indicating its superior capability in accurately identifying AAOCA. Conclusion We developed a high-performance AI tool for the fully automated detection of AAOCA using CCTA images. This tool can be used to evaluate large patient cohorts and detect AAOCA in extensive databases. Moreover, it can operate seamlessly alongside clinical assessments of CCTA scans, providing real-time alerts to medical personnel about potential AAOCA findings. This enhances diagnostic efficiency and supports early intervention in cardiovascular care.

The Countoscope: Measuring Self and Collective Dynamics without Trajectories

Driven by physical questions pertaining to quantifying particle dynamics, microscopy can now resolve complex systems at the single-particle level, from cellular organisms to individual ions. Yet, available analysis techniques face challenges reconstructing trajectories in dense and heterogeneous systems where accurately labeling particles is difficult. Furthermore, the inescapable finite field of view of experiments hinders the measurement of collective effects. Inspired by Smoluchowski, we introduce a broadly applicable analysis technique that probes dynamics of interacting particle suspensions based on a remarkably simple principle: counting particles in finite observation boxes. Using colloidal experiments, advanced simulations, and theory, we first demonstrate that statistical properties of fluctuating counts can be used to determine self-diffusion coefficients, so alleviating the hurdles associated with trajectory reconstruction. We also provide a recipe for practically extracting the diffusion coefficient from experimental data at variable particle densities, which is sensitive to steric and hydrodynamic interactions. Remarkably, by increasing the observation box size, counting naturally enables the study of collective dynamics in dense suspensions. Using our novel analysis of particle counts, we uncover a surprising enhancement of collective behavior due to hydrodynamics as well as a new length scale which can be connected with hyperuniform structure. Our counting framework, the “countoscope,” thus enables efficient measurements of self and collective dynamics in dense suspensions and opens the way to quantifying dynamics and identifying novel physical mechanisms in diverse complex systems where single particles can be resolved. Published by the American Physical Society 2024

Physicochemical properties of the confined hydrogen atom under dense semiclassical hydrogen plasma

Some fundamental quantities governing the physicochemical properties of the spherically confined (contained in a spherical box) hydrogen atom embedded in dense semiclassical hydrogen plasma have been investigated. These quantities specifically include the energy levels, wavefunctions, 2k-pole oscillator strength, 2k-pole polarizability, hyperfine spitting, effective pressure on the boundary of the confining surface. The effect of plasma is described by a pseudopotential which takes care of the collective effect and the quantum mechanical effects at short distances of the plasma particles by means of two adjustable parameters, namely, the screening parameter and the de Broglie wavelength. Energy eigenvalues of the atom for various box sizes and for different values of the plasma parameters are computed accurately within a variational framework by employing a large wavefunction which automatically takes care of the requisite boundary conditions. Convergence of the computed results is corroborated by increasing the number of terms in the wavefunction. Particular attention is paid on determining the critical size box for which all the bound states of the atom cease to exist. Based on the computed energies and the corresponding eigenfunctions, mean values of various powers of the radial coordinate, oscillator strengths, polarizability of various order, hyperfine splitting, and the effective pressure on the boundary have been evaluated. A comprehensive study is made on the changes of those quantities for varying box size and plasma parameters. Efforts are made to distinguish the changes arising out of the spatial confinement and the plasma confinement.

On the capabilities of k-ART over MD for the study of the kinetics of small point defect clusters in α-Fe

Molecular Dynamics simulations, while contributing to the understanding of the mechanisms of diffusion and interactions of point defects and their clusters, are inherently limited in their temporal scope (few nanoseconds). This constraint becomes particularly evident when studying the dynamics of vacancies at low temperatures, where their jump frequency is exceedingly low, posing challenges for accurate reproduction. Additionally, the size of the simulation box imposes constraints, influencing the representation of the system and potentially affecting the accuracy of results. A relatively new kinetic activation-relaxation technique (k-ART) efficiently resolves the limitations of MD simulations, such as computation time and system dimensionality, without the need for a priori knowledge of the simulated system. This technique enables simulations lasting up to several seconds and encompassing systems with higher dimensions. In this paper we check the validity of k-ART to reproduce accurately the migration mechanisms and energies of point defects and small clusters, previously obtained by MD and validated experimentally. We point out the advantages and difficulties of using AKMC with k-ART.

Antibiotic use in Barcelona in 2023 in primary care and the potential reduction by adjusting box sizes to current guidelines.

Prior research has not examined the size of antibiotic packages prescribed. We assessed 2023 prescription rates in Barcelona and the most prescribed presentations in pharmacies, while evaluating potential reductions if all amoxicillin and amoxicillin/clavulanate containers had 20 doses. Antibiotics prescribed by primary care doctors working for the Catalonian Health Institute in Barcelona in 2023 were analysed by calculating the defined daily doses per 1,000 inhabitants and day (DID). The observed prescribing rate was 8 DID, with penicillins accounting for 4.6 DID (57.2%). The most frequently prescribed antibiotics were amoxicillin and amoxicillin/clavulanate, making up 4.4 DID. If all the 30-dose presentations of amoxicillin and amoxicillin/clavulanate had been 20-dose containers, the total number of DIDs would have been 3.3, resulting in a reduction of 1.1 DID (25.4% less). Antibiotic prescribing rate in Barcelona was low. Aligning the dosage of antibiotics with established guidelines could further reduce antibiotic consumption.

On the Cosmic Variance of the Merger Rate Density of Binary Neutron Stars

The cosmic variance on the star formation history may lead to bias in the merger rate density estimation of binary neutron star (BNS) mergers by compact binary population synthesis. In this paper, we take advantage of the large box size of the Millennium Simulation combined with the semianalytic galaxy formation model GABE and the parameterized population binary star evolution model to examine how much effect the cosmic variance will introduce on the estimation of the merger rate density of BNS mergers. We find that for subbox sizes of 100 and 200 Mpc, the variance of merger rate density σ R /R at different redshifts is about 23%–35% and 13%–20%, respectively. On the one hand, as for the variance of the detection rate on BNS mergers with the current LIGO–Virgo–KAGRA (LVK) detector network, this value is very small at ≲10%, which indicates ignoring the cosmic variance is reasonable for estimating the merger rate density from current LVK observations. On the other hand, with next-generation gravitational wave detectors, it is possible to localize BNS mergers within subboxes possessing a length of 40 Mpc for a source redshift z s < 0.2. In such a small box, the cosmic variance of the merger rate density is significant, i.e., the value of σ R /R is about ∼55%. This hints that estimating the merger rate density of BNS in different sky areas may provide useful information on the cosmic variance.

Optimizing Object Detection Algorithms for Congenital Heart Diseases in Echocardiography: Exploring Bounding Box Sizes and Data Augmentation Techniques.

Congenital heart diseases (CHDs), particularly atrial and ventricular septal defects, pose significant health risks and common challenges in detection via echocardiography. Doctors often employ the cardiac structural information during the diagnostic process. However, prior CHD research has not determined the influence of including cardiac structural information during the labeling process and the application of data augmentation techniques. This study utilizes advanced artificial intelligence (AI)-driven object detection frameworks, specifically You Look Only Once (YOLO)v5, YOLOv7, and YOLOv9, to assess the impact of including cardiac structural information and data augmentation techniques on the identification of septal defects in echocardiographic images. The experimental results reveal that different labeling strategies substantially affect the performance of the detection models. Notably, adjustments in bounding box dimensions and the inclusion of cardiac structural details in the annotations are key factors influencing the accuracy of the model. The application of deep learning techniques in echocardiography enhances the precision of detecting septal heart defects. This study confirms that careful annotation of imaging data is crucial for optimizing the performance of object detection algorithms in medical imaging. These findings suggest potential pathways for refining AI applications in diagnostic cardiology studies.

EXPRESS: A Machine Learning Approach to Solve the E-commerce Box-Sizing Problem

E-commerce packages are notorious for their inefficient usage of space. More than one-quarter volume of a typical e-commerce package comprises air and filler material. The inefficient usage of space significantly reduces the transportation and distribution capacity increasing the operational costs. Therefore, designing an optimal set of packaging box sizes is crucial for improving efficiency. We present the first learning-based framework to determine the optimal packaging box sizes. In particular, we propose a three-stage optimization framework that combines unsupervised learning, reinforcement learning, and tree search to design box sizes. The package optimization problem is formulated into a sequential decision-making task called the box-sizing game. A neural network agent is then designed to play the game and learn heuristic rules to solve the problem. In addition, a tree-search operator is developed to improve the performance of the learned networks. When benchmarked with company-based optimization formulation and two alternate optimization models, we find that our ML-based approach can effectively solve large-scale problems within a stipulated time. We evaluated our model on real-world datasets supplied by a large e-commerce platform. The framework is currently adopted by a large e-commerce company across its 28 fulfillment centers, which is estimated to save the company about 7.1 million USD annually. In addition, it is estimated that paper consumption will be reduced by 2080 metric tons and greenhouse gas emissions by 1960 metric tons annually. The presented optimization framework serves as a decision support tool for designing packaging boxes at large e-commerce warehouses.

Revisiting evolution of domain walls and their gravitational radiation with CosmoLattice

Employing the publicly available 𝒞osmoℒattice code, we conduct numerical simulations of a domain wall network and the resulting gravitational waves (GWs) in a radiation-dominated Universe in the Z2-symmetric scalar field model. In particular, the domain wall evolution is investigated in detail both before and after reaching the scaling regime, using the combination of numerical and theoretical methods. We demonstrate that the total area of closed walls is negligible compared to that of a single long wall stretching throughout the simulation box. Therefore, the closed walls are unlikely to have a significant impact on the overall network evolution. This is in contrast with the case of cosmic strings, where formation of loops is crucial for maintaining the system in the scaling regime. To obtain the GW spectrum, we develop a technique that separates physical effects from numerical artefacts arising due to finite box size and non-zero lattice spacing. Our results on the GW spectrum agree well with refs. [29,30], which use different codes. Notably, we observe a peak at the Hubble scale, an exponential falloff at scales shorter than the wall width, and a plateau/bump at intermediate scales. We also study sensitivity of obtained results on the choice of initial conditions. We find that different types of initial conditions lead to qualitatively similar domain wall evolution in the scaling regime, but with important variations translating into different intensities of GWs.

Integration of LIME Explainable AI to Enhance Interpretability of Deep Learning Models in Box Palette Classification

In the food production system, manual box palette arrangement often encounters errors, such as incorrect stacking patterns, mismatched box quantities, and improper mixing of product variants. This issue occurs in a soy sauce production company with limited infrastructure, leading to the mixing of two product variants with the same box size in a single production line. This results in disruptions to the production flow and significant potential losses. This study proposes a solution using deep learning to detect box palette arrangement patterns. The Convolutional Neural Network (CNN) method is chosen because it has proven effective in image classification. Additionally, this study implements Explainable AI (XAI) to provide explanations related to classification results, increasing user confidence in the system. The Local Interpretable Model-agnostic Explanations (LIME) technique will be used to provide interpretations. This research produces the development of a deep learning model to classify box palette arrangements. Furthermore, the implementation of LIME in this study successfully provides interpretations of the model's predictions. The result is evident in the result where it shows that MobileNetV2 give an F1-Score of 100%, and LIME fidelity score of 0.2 and stability score of 0.2.

Asimulation: Domain formation and impact on observables in resolved cosmological simulations of the (a)symmetron

The symmetron is a dark energy and dark matter candidate that forms topological defects in the late-time universe and holds the promise of resolving some of the cosmological tensions. We performed high-resolution simulations of the dynamical and non-linear (a)symmetron using the recently developed relativistic N-body code asevolution. By extensively testing the temporal and spatial convergence of domain decompositioning and domain wall stability, we determined criteria and physical intuition for the convergence. We applied the resolution criteria to run five high-resolution simulations with 12803 grids and a box size of 500 Mpc h−1 of the (a)symmetron. We considered the behaviour of the scalar field and the domain walls in each scenario. We find the effect on the matter power spectra, the HMFs, and observables computed over the past light cone of an observer, such as the integrated Sachs-Wolfe and non-linear Rees-Sciama effect and the lensing, compared to ΛCDM. We show local oscillations of the fifth force strength and the formation of planar structures in the density field. The dynamics of the field was visualised in animations with high resolution in time. The simulation code is made publicly available.

Evaluation of the use of box size priors for 6D plane segment tracking from point clouds with applications in cargo packing

This paper addresses the problem of 6D pose tracking of plane segments from point clouds acquired from a mobile camera. This is motivated by manual packing operations, where an opportunity exists to enhance performance, aiding operators with instructions based on augmented reality. The approach uses as input point clouds, by its advantages for extracting geometric information relevant to estimating the 6D pose of rigid objects. The proposed algorithm begins with a RANSAC fitting stage on the raw point cloud. It then implements strategies to compute the 2D size and 6D pose of plane segments from geometric analysis of the fitted point cloud. Redundant detections are combined using a new quality factor that predicts point cloud mapping density and allows the selection of the most accurate detection. The algorithm is designed for dynamic scenes, employing a novel particle concept in the point cloud space to track detections’ validity over time. A variant of the algorithm uses box size priors (available in most packing operations) to filter out irrelevant detections. The impact of this prior knowledge is evaluated through an experimental design that compares the performance of a plane segment tracking system, considering variations in the tracking algorithm and camera speed (onboard the packing operator). The tracking algorithm varies at two levels: algorithm (Awpk\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$A_{wpk}$$\\end{document}), which integrates prior knowledge of box sizes, and algorithm (Awoutpk\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$A_{woutpk}$$\\end{document}), which assumes ignorance of box properties. Camera speed is evaluated at low and high speeds. Results indicate increments in the precision and F1-score associated with using the Awpk\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$A_{wpk}$$\\end{document} algorithm and consistent performance across both velocities. These results confirm the enhancement of the performance of a tracking system in a real-life and complex scenario by including previous knowledge of the elements in the scene. The proposed algorithm is limited to tracking plane segments of boxes fully supported on surfaces parallel to the ground plane and not stacked. Future works are proposed to include strategies to resolve this limitation.

Object Detection and Monocular Stable Distance Estimation for Road Environments: A Fusion Architecture Using YOLO-RedeCa and Abnormal Jumping Change Filter

Enabling rapid and accurate comprehensive environmental perception for vehicles poses a major challenge. Object detection and monocular distance estimation are the two main technologies, though they are often used separately. Thus, it is necessary to strengthen and optimize the interaction between them. Vehicle motion or object occlusions can cause sudden variations in the positions or sizes of detection boxes within temporal data, leading to fluctuations in distance estimates. So, we propose a method to integrate a detector based on YOLOv5-RedeCa, a Bot-Sort tracker and an anomaly jumping change filter. This combination allows for more accurate detection and tracking of objects. The anomaly jump filter smooths distance variations caused by sudden changes in detection box sizes. Our method increases accuracy while reducing computational demands, showing outstanding performance on several datasets. Notably, on the KITTI dataset, the standard deviation of the continuous ranging results remains consistently low, especially in scenarios with multiple object occlusions or disappearances. These results validate our method’s effectiveness and precision in managing dual tasks.

Intra-Frame Graph Structure and Inter-Frame Bipartite Graph Matching with ReID-Based Occlusion Resilience for Point Cloud Multi-Object Tracking

Three-dimensional multi-object tracking (MOT) using lidar point cloud data is crucial for applications in autonomous driving, smart cities, and robotic navigation. It involves identifying objects in point cloud sequence data and consistently assigning unique identities to them throughout the sequence. Occlusions can lead to missed detections, resulting in incorrect data associations and ID switches. To address these challenges, we propose a novel point cloud multi-object tracker called GBRTracker. Our method integrates an intra-frame graph structure into the backbone to extract and aggregate spatial neighborhood node features, significantly reducing detection misses. We construct an inter-frame bipartite graph for data association and design a sophisticated cost matrix based on the center, box size, velocity, and heading angle. Using a minimum-cost flow algorithm to achieve globally optimal matching, thereby reducing ID switches. For unmatched detections, we design a motion-based re-identification (ReID) feature embedding module, which uses velocity and the heading angle to calculate similarity and association probability, reconnecting them with their corresponding trajectory IDs or initializing new tracks. Our method maintains high accuracy and reliability, significantly reducing ID switches and trajectory fragmentation, even in challenging scenarios. We validate the effectiveness of GBRTracker through comparative and ablation experiments on the NuScenes and Waymo Open Datasets, demonstrating its superiority over state-of-the-art methods.

HP-YOLOv8: High-Precision Small Object Detection Algorithm for Remote Sensing Images.

YOLOv8, as an efficient object detection method, can swiftly and precisely identify objects within images. However, traditional algorithms encounter difficulties when detecting small objects in remote sensing images, such as missing information, background noise, and interactions among multiple objects in complex scenes, which may affect performance. To tackle these challenges, we propose an enhanced algorithm optimized for detecting small objects in remote sensing images, named HP-YOLOv8. Firstly, we design the C2f-D-Mixer (C2f-DM) module as a replacement for the original C2f module. This module integrates both local and global information, significantly improving the ability to detect features of small objects. Secondly, we introduce a feature fusion technique based on attention mechanisms, named Bi-Level Routing Attention in Gated Feature Pyramid Network (BGFPN). This technique utilizes an efficient feature aggregation network and reparameterization technology to optimize information interaction between different scale feature maps, and through the Bi-Level Routing Attention (BRA) mechanism, it effectively captures critical feature information of small objects. Finally, we propose the Shape Mean Perpendicular Distance Intersection over Union (SMPDIoU) loss function. The method comprehensively considers the shape and size of detection boxes, enhances the model's focus on the attributes of detection boxes, and provides a more accurate bounding box regression loss calculation method. To demonstrate our approach's efficacy, we conducted comprehensive experiments across the RSOD, NWPU VHR-10, and VisDrone2019 datasets. The experimental results show that the HP-YOLOv8 achieves 95.11%, 93.05%, and 53.49% in the mAP@0.5 metric, and 72.03%, 65.37%, and 38.91% in the more stringent mAP@0.5:0.95 metric, respectively.

FireYOLO-Lite: Lightweight Forest Fire Detection Network with Wide-Field Multi-Scale Attention Mechanism

A lightweight forest fire detection model based on YOLOv8 is proposed in this paper in response to the problems existing in traditional sensors for forest fire detection. The performance of traditional sensors is easily constrained by hardware computing power, and their adaptability in different environments needs improvement. To balance the accuracy and speed of fire detection, the GhostNetV2 lightweight network is adopted to replace the backbone network for feature extraction of YOLOv8. The Ghost module is utilized to replace traditional convolution operations, conducting feature extraction independently in different dimensional channels, significantly reducing the complexity of the model while maintaining excellent performance. Additionally, an improved CPDCA channel priority attention mechanism is proposed, which extracts spatial features through dilated convolution, thereby reducing computational overhead and enabling the model to focus more on fire targets, achieving more accurate detection. In response to the problem of small targets in fire detection, the Inner IoU loss function is introduced. By adjusting the size of the auxiliary bounding boxes, this function effectively enhances the convergence effect of small target detection, further reducing missed detections, and improving overall detection accuracy. Experimental results indicate that, compared with traditional methods, the algorithm proposed in this paper significantly improves the average precision and FPS of fire detection while maintaining a smaller model size. Through experimental analysis, compared with YOLOv3-tiny, the average precision increased by 5.9% and the frame rate reached 285.3 FPS when the model size was only 4.9 M; compared with Shufflenet, the average precision increased by 2.9%, and the inference speed tripled. Additionally, the algorithm effectively addresses false positives, such as cloud and reflective light, further enhancing the detection of small targets and reducing missed detections.