Centroid Position Research Articles

Single-Shot, Monochrome, Spatial Pixel-Encoded, Structured Light System for Determining Surface Orientations

This study introduces a technique for determining surface orientations by projecting a monochrome, spatial pixel-encoded pattern and calculating the surface normals from single-shot measurement. Our method differs from traditional methods, such as shape from shading and shape from texture, in that it does not require relating the local surface orientations of adjacent points. We propose a multi-resolution system incorporating symbols varying in sizes from 8 × 8, 10 × 10, 12 × 12, 14 × 14, and 16 × 16 pixels. Compared to previous methods, we have achieved a denser reconstruction and obtained a 5.2 mm resolution using an 8 × 8 pattern at a depth of 110 cm. Unlike previous methods, which used local point orientations of grid intersection and multiple colors, we have used the monochrome pattern and deterministic centroid positions to compute the unit vector or direction vector between the neighboring symbols. The light plane intersections are used to calculate the tangent vectors on the surface. Surface normals are determined by the cross-product of two tangent vectors on the surface. A real experiment was conducted to measure simple plane surfaces, circular surfaces, and complex sculptures. The results show that the process of calculating surface normals is fast and reliable, and we have computed 1654 surface normals in 29.4 milliseconds for complex surfaces such as sculptures.

Multi-body dynamics modeling and energy consumption optimization of electromagnetic mechanical fully variable valve system

Electromagnetic Mechanical Variable Valve Actuation (EMVVA) technology has the capability to improve engine performance and decrease pollutant emissions. A new type of EMVVA composed of electromagnetic driver and mechanical transmission is proposed to realize the flexible adjustment of valve parameters whether the valve is open or closed process. The geometric mathematical model of the mechanical transmission is developed based on the motion laws of the valve and the geometrical construction of the mechanical transmission, and the conjugate cam curve is solved. A multi-body dynamic model is constructed to calculate the driving torque and energy consumption needed by the mechanical transmission based on the mass, rotational inertia, centroid position of the parts, combining the normal contact force model and the friction model of the clearance contact pair at the same time. Based on the geometric model and multi-body dynamics model, the test platform is established. The test results demonstrated that within the specified valve lift, EMVVA system could accomplish variable valve lift, variable valve timing, and variable valve duration at maximum lift. The maximum timing error does not exceed 1° crank angle, and the maximum valve lift error does not exceed 0.25 mm at 11 mm valve lift. In addition, the error of driving torque between the multi-body dynamics model and the test is less than 0.3 [Formula: see text].Low energy optimization of mechanical transmission was completed using the NSGA-II method and multi-body model. According to the results of the optimization, energy consumption was reduced by 14%, and the peak driving torque was decreased by 54.3%.

Experimental Method for Measuring Cohesion of Regolith via Electrostatic Lofting

The hypothesized electrostatic lofting of individual regolith grains on the Moon and asteroids has been investigated extensively in laboratory studies. Cohesion may dominate how regolith behaves on these small, airless bodies, yet the magnitude of this force remains uncertain. We induce the electrostatic detachment of dust as a mechanism to break cohesive bonds between individual zirconia-silica microspheres in order to measure the interparticle cohesive force between them, likely dominated by capillary bridges. A high-speed camera imaged centroid positions of the lofted microspheres over time. Using the centroids from the initial detachment, we numerically calculated initial accelerations to solve for the cohesion that had been restraining the microspheres. Unexpectedly, the electrostatic lofting of clumps of particles was observed and experimental results showed that clumps were a nonnegligible portion of the lofted object population.

Experimental study on the influence of centroid position of wind turbine section on flutter

Aiming at the large aspect ratio characteristics of large wind turbines, it is easy to cause irreversible damage due to flutter during operation. A two-degree-of-freedom (plunge and pitch) flutter test bench was built using the time domain and frequency domain analysis methods of dynamic signals. The influence of different centroid positions on the flutter boundary was studied. The test shows that the closer the centroid position is to the leading edge of the wing segment, the better the aeroelastic stability of the wing segment is. Under the linear condition, the forward movement of the centroid position has a more significant influence on the flutter. In addition, the main reason for the wing flutter is related to the decrease of net damping and the coupling of the aeroelastic natural frequency of the pitching and plunging motions. The pitch motion is dominant in the two-degree-of-freedom motion. The farther the centre of mass is from the torsion axis, the greater the pitch and plunge motion displacement. The pitch motion has a more significant impact on the system than the plunge motion. Therefore, the study of flutter suppression should focus on pitch motion.

A simple subtropical high‐pressure system index over the South Atlantic

AbstractThis research introduces a novel index for the South Atlantic High Pressure (SAHP) system to enhance understanding of regional climate variability and change. Subtropical highs significantly influence regional climates, yet comprehensive indices to measure their behaviours are lacking. Utilizing ERA5 reanalysis data from 1940 to 2023, the proposed index estimates a weighted centroid of the area surrounding the maximum sea level pressure within a 3 hPa range. This method ensures robustness and flexibility in contiguous area estimation specific to subtropical high events. Results showed the index effectively reflects the position and intensity of the SAHP. The study reveals that latitudinal variability of the SAHP has a strong unimodal structure, whereas longitudinal variability exhibits a bimodal structure. Seasonal patterns of the index show noticeable changes, with winter (JJA) and spring (SON) months having relatively high index values compared to summer (DJF) and autumn (MAM) months, underscoring the intra‐annual variability of the SAHP index. During ENSO events, the mean centroid position of the SAHP shifts significantly, moving westwards and polewards during El Niño and showing greater stability during La Niña. The index, with minimal computation requirements and flexibility, can be applied across diverse datasets, aiding in the assessment of future subtropical high changes.

Data-driven Bayesian Gaussian mixture optimized anchor box model for accurate and efficient detection of green citrus

The demand for harvesting green citrus has increased due to the rapid growth of the green citrus industry. However, real-time detection of green citrus is challenging because of its high visual similarity to the background. To address this issue, a new model called EDGC-YOLO (Efficient Detection of Green Citrus based on the YOLO) has been developed, focusing on detecting green citrus effectively and accurately. Specifically, the proposed model uses a data-driven approach that employs the highest confidence level from the aspect ratio of unobstructed citrus fruits to infer anchor boxes of obscured citrus. This method ensures a high degree of coincidence with bounding boxes based on empirical annotation. Furthermore, analyzing the distribution range of various citrus annotation categories within the dataset and the incorporating the highest confidence level of the normal distribution regarding aspect ratios, the initial cluster centers of the BGMM (Bayesian Gaussian Mixture Model) are dynamically adjusted. This strategy enables the anchor boxes to be inferred from the best suited for detecting green citrus. These improved anchor boxes provide a more accurate and effective solution for centroid positioning of obscured green citrus, significantly enhancing the entire network model’s detection accuracy and lightweight efficiency. To further improve detection accuracy and reduce model size, this study integrated the Refined-EfficientNetV2 network as the backbone, enhanced with a Convolutional Block Attention Module (CBAM) for better feature extraction. This integration allows the model to effectively capture relevant channel information and spatial features, increasing the feature extraction capability for green citrus images. Experimental results demonstrate the model’s superior performance: parameters reduced to 4.52 million, computational demand to 7.8 GFLOPs (64.4 % and 49.4 % of the original model, respectively), and model size decreased to 9.4 MB (65.3 % of the original). Additionally, the optimized model improved Precision (P), Recall (R), and mean Average Precision (mAP) by 0.5 %, 1.6 %, and 3.0 %, respectively, compared to the original model. The proposed model achieves higher detection accuracy with a smaller model size, providing theoretical support for the harvesting decision-making of green citrus harvesting robots.

Potential anatomical triggers for plan adaptation of cervical cancer external beam radiotherapy.

This study aimed to identify potential anatomical variation triggers using magnetic resonance imaging for plan adaption of cervical cancer patients to ensure dose requirements were met over an external beam radiotherapy course. Magnetic resonance images (MRIs) acquired before and during treatment were rigidly registered to a pre-treatment computerised tomography (CT) image for 11 retrospective cervix cancer datasets. Target volumes (TVs) and organs at risk (OARs) were delineated on both MRIs and propagated onto the CT. Treatment plans were generated based on the pre-treatment contours and applied to the mid-treatment contours. Anatomical and dosimetric changes between each timepoint were assessed. The anatomical changes included the change in centroid position and volume size. Dosimetric changes included the V30Gy and V40Gy for the OARs, and V95%, V100%, D95% and D98% for the TVs. Correlation with dosimetric and anatomical changes were assessed to determine potential replan triggers. Changes in the bowel volume and position in the superior-inferior direction, and the high-risk CTV anterior posterior position were highly correlated with a change in dose to the bowel and target, respectively. Hence changes in bowel and high-risk CTV could be used as a potential replan triggers.

Optical tracking and size estimation of a moving object via time-division multiplexing ghost imaging

Tracking and imaging a moving target via single-pixel imaging requires completion within a short period. Therefore, a low sampling rate is necessary to prevent imaging failure. This paper proposes a method utilizing time-division multiplexing ghost imaging, employing geometric moment patterns and Fourier fringe patterns to obtain real-time target trajectories and Fourier coefficient slices of the scene. After calculating the target’s displacement relative to the initial moment, we perform motion compensation on the slices to gradually estimate the size of the target. The method has been validated as effective and adaptive through optical experiments, requiring only a small amount of sampling relative to the image resolution. Our approach can achieve real-time tracking and estimate the region occupied by the object in the scene and the centroid position relative to the object. It considers the sparsity of objects in space, offering a promising solution for future low-sampling-rate and high-resolution single-pixel imaging of a moving target.

The Effect of Model Input Uncertainty on the Simulation of Typical Pollutant Transport in the Coastal Waters of China

Route planning to evade potential pollution holds critical importance for aquaculture vessels. This study establishes a fish-feed pollutant drift model based on the Lagrangian particle tracking algorithm and designs four sets of sensitivity experiments in the East China Sea. The research investigates the impact of model input uncertainties on the drift trajectory, centroid position, and sweeping area of the fish-feed pollutants. Numerical results indicate that the uncertainty in the background flow field significantly affects the uncertainty in the centroid position and sweeping area in the numerical simulations. Specifically, when a 35% random error is added to the background flow field, the centroid shift distance reaches its maximum, and the sweeping area also attains its largest value. The uncertainty in the background wind field affects the centroid position of particles but to a much lesser extent compared to the background flow field. When considering only the uncertainty of the background wind field, the sweeping area does not significantly differ from the control experiment as the uncertainty of the background wind field increases. The initial release position has little effect on the drift direction of the fish-feed pollutants but does affect the drift distance; it has minimal impact on the trajectory but significantly affects the final position of the pollutant centroid. By analyzing the model uncertainties, this study reveals the key factors influencing the drift of fish-feed pollutants. This information is crucial for aquaculture vessels in planning routes, considering environmental factors, and reducing potential pollution risks.

The vertex centroid of a Van Aubel result involving similar quadrilaterals and its further generalisation

This paper explores the position of the vertex centroid for a generalisation of Van Aubel’s theorem: specifically we look at what happens to the vertex centroid when directly similar quadrilaterals are placed on the sides of an arbitrary quadrilateral. After giving a simple proof that the position of the vertex centroid remains unchanged, the result is further generalised to directly similar triangles (or other directly similar shapes) on the sides of polygons using vectors. Not only are the results mathematically interesting, but can also provide an appropriate classroom opportunity for dynamic geometry exploration, and to build 2D models with clay and drinking straws (or thin wire) to illustrate and check the theoretical solutions.

Hierarchical-agglomerative clustering analysis of geomorphic features applied to tectonic investigation of terrestrial planets: An example from Claritas Fossae, Mars

A hierarchical-agglomerative clustering workflow is proposed to investigate complex, tectonically controlled regions on planetary surfaces. This algorithm comprises two primary steps: (i) calculating the dissimilarity between each object and (ii) grouping objects using a hierarchical clustering method. The efficiency of the workflow hinges on two critical parameters that require careful selection: the attributes for grouping objects and the number of clusters for interpretation. We applied this approach to tectonic lineaments in the Claritas Fossae (CF) region on Mars, a complex area significantly shaped by tectonic activity. Our analysis considered three attributes of these lineaments: (i) azimuthal direction, (ii) length, and (iii) centroid position. The optimal number of clusters was determined using the Silhouette index (S), which assesses the robustness of the clustering results. Our objectives were twofold: (i) to replicate the distribution of lineament sets obtained from classical geostatistical analysis of their azimuthal orientation and (ii) to refine the subdivision by varying combinations of the three attributes. Our results provide crucial insights into the geo-tectonic evolution of CF, supporting the hypothesis of a tectonic evolution characterised by a polyphase history. In addition, we demonstrate how the method is capable of effectively identified areas with varying intensities of brittle deformation.

Research on the Influence of Radial Variation of Centroid on the Motion of Spherical Robot

Through the pendulum mechanism inside the spherical shell, the centroid can be varied circumferentially, enabling the spherical robot to achieve omnidirectional flexible movement. Additionally, the radial variation ability of the centroid enables spherical robots to adopt two distinct driving modes: the traditional lower pendulum driving mode and the inverted pendulum driving mode. There are two manifestations of radial variation in the centroid: having different radial positions of the centroid and achieving radial movement of the centroid. Focusing on these two manifestations, experimental data are obtained through different motion velocities and different motion slopes to conduct research on the influence of radial variation in the centroid on the motion of spherical robots. Based on the experimental data, multiple indicators are analyzed, including response speed, convergence speed, stability, and overshoot, as well as steering ability, climbing ability, and output power. The impact of the radial variation ability of the centroid on the control performance, locomotion capability, and energy consumption of spherical robots is summarized, and the correlation model relating the motion requirements to the radial position of the centroid is established, providing a theoretical basis for the selection of driving modes and centroid positions for spherical robots facing complex task requirements.

Toward an integrated multi-gigahertz ionizing particle diagnostic.

Needs arising at both current and future accelerator facilities call for the development of radiation-hardened position-sensing diagnostics that can operate with multi-GHz repetition rates. Such instruments are likely to also have applications in the diagnosis of rapid plasma behavior. Building on the recent work of our Advanced Accelerator Diagnostics Collaboration, we are exploring the development of integrated multi-GHz ionizing particle detection systems based on chemical-vapor deposition diamond sensors, with the initial goal of producing a quadrant detector that can determine the intensity and centroid position of a particle beam at a repetition rate between 5 and 10GHz. Results from our initial high-speed characterization work are presented, including those from a single-channel sensor with a GHz response. Approaches to achieving multi-GHz (5-10GHz) rate capability, including the design of a dedicated Application Specific Integrated Circuit and the use of 3D RF-solver computer aided design software, are presented and discussed in more detail. 3D RF simulations suggest clean pulses of duration less than 250ps (FWHM less than 125ps) can be achieved with the approaches developed by this work.

A de-texturing model for enhancing the accuracy of centroid positioning

In tasks guided by microvision, extracting centroids is a common method for positioning, which is negatively affected by texture. Here, an attention-related de-texturing model is proposed to eliminate the texture of microparts and preserve accurate edges. A network with an attention module called De-texturing Net is built, in which both the transformer and channel attention modules are included. Considering the importance of texture, the additional factor in loss function is constructed based on the Gram matrix difference between target images and generated images. Results show that De-texturing Net can generate de-texturized images with high Peak Signal to Noise Ratio/SSIM, indicating the similarity between de-texturized and target images. Moreover, for the centroid positioning, the error in de-texturized images is significantly lower than the error in original images. This study helps improve the accuracy of centroid positioning due to the de-texturized images with accurate edges.

Convolutional neural network for improved event-based Shack-Hartmann wavefront reconstruction

Shack-Hartmann wavefront sensing is a technique for measuring wavefront aberrations, whose use in adaptive optics relies on fast position tracking of an array of spots. These sensors conventionally use frame-based cameras operating at a fixed sampling rate to report pixel intensities, even though only a fraction of the pixels have signal. Prior in-lab experiments have shown feasibility of event-based cameras for Shack-Hartmann wavefront sensing (SHWFS), asynchronously reporting the spot locations as log intensity changes at a microsecond time scale. In our work, we propose a convolutional neural network (CNN) called event-based wavefront network (EBWFNet) that achieves highly accurate estimation of the spot centroid position in real time. We developed a custom Shack-Hartmann wavefront sensing hardware with a common aperture for the synchronized frame- and event-based cameras so that spot centroid locations computed from the frame-based camera may be used to train/test the event-CNN-based centroid position estimation method in an unsupervised manner. Field testing with this hardware allows us to conclude that the proposed EBWFNet achieves sub-pixel accuracy in real-world scenarios with substantial improvement over the state-of-the-art event-based SHWFS. An ablation study reveals the impact of data processing, CNN components, and training cost function; and an unoptimized MATLAB implementation is shown to run faster than 800 Hz on a single GPU.

Imaging a Large Coronal Loop Using Type U Solar Radio Burst Interferometry

Solar radio U-bursts are generated by electron beams traveling along closed magnetic loops in the solar corona. Low-frequency (<100 MHz) U-bursts serve as powerful diagnostic tools for studying large-sized coronal loops that extend into the middle corona. However, the positive frequency drift component (descending leg) of U-bursts has received less attention in previous studies, as the descending radio flux is weak. In this study, we utilized LOFAR interferometric solar imaging data from a U-burst that has a significant descending leg component, observed between 10 and 90 MHz on 2020 June 5th. By analyzing the radio source centroid positions, we determined the beam velocities and physical parameters of a large coronal magnetic loop that reached just about 1.3 R ⊙ in altitude. At this altitude, we found the plasma temperature to be around 1.1 MK, the plasma pressure around 0.20 mdyn, cm−2, and the minimum magnetic field strength around 0.07 G. The similarity in physical properties determined from the image suggests a symmetric loop. The average electron beam velocity on the ascending leg was found to be 0.21c, while it was 0.14c on the descending leg. This apparent deceleration is attributed to a decrease in the range of electron energies that resonate with Langmuir waves, likely due to the positive background plasma density gradient along the downward loop leg.

A Multi-Target Identification and Positioning System Method for Tomato Plants Based on VGG16-UNet Model

The axillary buds that grow between the main and lateral branches of tomato plants waste nutrients and lead to a decrease in yield, necessitating regular removal. Currently, these buds are removed manually, which requires substantial manpower and incurs high production costs, particularly on a large scale. Replacing manual labor with robots can lead to cost reduction. However, a critical challenge is the accurate multi-target identification of tomato plants and precise positioning for axillary bud removal. Therefore, this paper proposes a multi-target identification and localization method for tomato plants based on the VGG16-UNet model. The average intersection and pixel accuracies of the VGG16-UNet model after introducing the pretrained weights were 85.33% and 92.47%, respectively, which were 5.02% and 4.08% higher than those of the VGG16-UNet without pretrained weights, achieving the identification of main branches, side branches, and axillary bud regions. Then, based on the multi-objective segmentation of the tomato plants in the VGG16-UNet model, the regions of the axillary buds in the tomato plants were identified by HSV color space conversion and color threshold range selection. Morphological dilation and erosion operations were used to remove noise and connect adjacent regions of the same target. The endpoints and centroids of the axillary buds were identified using the feature point extraction algorithm. The left and right positions of the axillary buds were judged by the relationship between the position of the axillary bud centroid and the position of the main branch. Finally, the coordinate parameters of the axillary bud removal points were calculated using the feature points to determine the relationship between the position of the axillary bud and the position of the branch. Experimental results showed that the average accuracy of the axillary bud pruning point recognition was 85.5%.

LollipopE: Bi-centered lollipop embedding for complex logic query on knowledge graph

Answering complex First-Order Logic (FOL) query plays a vital role in multi-hop knowledge graph (KG) reasoning. Geometric methods have emerged as a promising category of approaches in this context. However, existing best-performing geometric query embedding (QE) model is still up against three-fold potential problems: (i) underutilization of embedding space, (ii) overreliance on angle information, (iii) uncaptured hierarchy structure. To bridge the gap, we propose a lollipop-like bi-centered query embedding method named LollipopE. To fully utilize embedding space, LollipopE employs learnable centroid positions to represent multiple entities distributed along the same axis. To address the potential overreliance on angular metrics, we design an angular-based and centroid-based metric. This involves calculating both an angular distance and a centroid-based geodesic distance, which empowers the model to make more informed selections of relevant answers from a wider perspective. To effectively capture the hierarchical relationships among entities within the KG, we incorporate dynamic moduli, which allows for the representation of the hierarchical structure among entities. Extensive experiments demonstrate that LollipopE surpasses the state-of-the-art geometric methods. Especially, on more hierarchical datasets, LollipopE achieves the most significant improvement.

Abstract 286: Multiplex imaging of human pancreatic tumors reveal novel dynamics of dysplastic stem cells, remodeling of associated fibroblasts and cell-to-cell spatial dependencies

Abstract This study aims to investigate human normal pancreatic tissue, pancreatic precursor lesions, and invasive pancreatic ductal adenocarcinoma to provide the most extensive spatial characterization as disease progress by mapping dysplastic stem cells (DSCs) - a reservoir of cancer cells - and fibroblast populations which modulate the tumor microenvironment and participate in disease progression. We performed 12 cycles of multiplex immunofluorescence using 26 validated antibodies covering fibroblasts [CXCL12, αSMA, FAP, CD74, CD105, PDGFRA, COL1A1], DSCs diversity [CD44v9, CD133, CD166, SPP1, S100P], and immune populations, onto tissue samples from 75 patients screened for normal, low grade, high grade dysplasia, and invasive cancer. We used our codes in Python, R, and QuPath 0.4.3. The pipeline includes shading correction, whole slide image registration (at subcellular level), region of interest and stromal mask annotations by a pathologist [normal duct, acinar to ductal metaplasia (ADM), low grade, high grade, invasive], automatic cell detection, and signal extraction. We developed a novel Python code for cell-to-cell spatial dependency analysis by using the x,y centroid positions of one cell from the tumor to other cells from the stroma mask based on pre-defined distances. For this abstract, we interrogated 264,732 cells and observed SPP1+, S100P+ DSCs expansion from low grade to high grade dysplasia, whereas CD133+ and CD166+ populations contract (p&amp;lt;0.0001). S100P+ DSCs gradually expand as disease progress reaching 74.13% of invasive cells being positive. CD44v9 follows dynamic changes variations, being expressed in 59.6% of cells undergoing ADM, dropping to 5.4% in dysplasia, and re-increasing to 25.4% in invasive cancer (all p&amp;lt;0.0001). Focusing on stromal composition, CXCL12+, CD105+, αSMA+, CD74+ fibroblasts expand whereas FAP+ fibroblasts contract from low- to high grade (p-values &amp;lt;0.0001). COL1A1 expression increases at each step of the normal duct - ADM - low grade - high grade - invasive cancer sequence (all p&amp;lt;0.0001). The pattern of CD44v9 tumor population showed dynamic alterations with the CD105 fibroblast population, as CD44v9 and CD105 are highly expressed in coordination during ADM. However, in dysplastic lesions and invasive cancer, mutual exclusivity between the tumor CD44v9 populations and the underlying CD105 fibroblast populations is observed, confirmed by our spatial algorithm within 50 μm of cell-to-cells distance (p&amp;lt;0.0001). In tumors, when CD44v9 tumor cell populations are expressed, the adjacent fibroblasts were CD105 negative. Multiplex imaging reveals dysplastic stem cells and fibroblasts dynamics in low- to high grade transition. The high-throughput capabilities of this method lead to the discovery of the CD44v9 tumor and CD105 fibroblast dependencies in pancreatic tumors. Citation Format: Maëlle Batardière, Shoukari Ayman, Camille Beaussier, Jumanah Baig, Zean Ghanmeh, Melisa Farias-Gonzalez, Jahg Wong, Alexandre Archambault-Marsan, Louise Rousseau, Simon Turcotte, Anna Means, Marcus Tan, Kathleen DelGiorno, Elham Dianati, Vincent Quoc-Huy Trinh. Multiplex imaging of human pancreatic tumors reveal novel dynamics of dysplastic stem cells, remodeling of associated fibroblasts and cell-to-cell spatial dependencies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 286.

Quantitative analysis of sintered NdFeB backscattered electron images based on a general large model

The macroscopic performance of magnets is determined by their microscopic structure, quantifying the microscopic image of magnets is of great importance for studying its performance. Backscattered electron images of sintered NdFeB magnets contain information about the size, morphology, and distribution of grains and the grain boundary phases. Traditional methods for quantifying images involve labor-intensive manual measurements, digital image processing with complex contour extraction algorithms, and convolutional neural network algorithms that require extensive image data labeling. In this study, we introduced a general vision large model called Segment Anything Model(SAM) for image segmentation. SAM enables rapid and accurate segmentation of grains and grain boundary phases without the need for complex algorithms and tedious data labeling. From the segmented mask images, we extracted various data related to the performance of magnets, including the centroid positions, perimeter, area, sphericity, roughness, and principal axis directions of all grains. We also obtained information on the distances and angles between adjacent grains and the relevant parameters affecting magnet performance, such as the number and volume of grain boundary phases. We conducted comprehensive quantification of the backscattered images for three different magnets and provided reasonable explanations for the differences in magnet performance. This model offers superior speed and accuracy in image quantification compared to traditional algorithms and can be used for the rapid analysis of large datasets. It represents an essential method and trend for the quantification of image data in the future.