Contour Data Research Articles

Quantification of facial symmetry in orthognathic surgery: A novel approach integrating 3D contour maps and hyper-dimensional computing

This study aimed to enhance the evaluation of facial symmetry crucial for planning and assessing outcomes of orthognathic surgery (OGS). An innovative approach combining three-dimensional (3D) facial contour lines with hyperdimensional (HD) computing was developed for this purpose. Data were collected using 3D cone beam computed tomography (CBCT) at Chang Gung Memorial Hospital from 2016 to 2021. A comprehensive dataset was compiled, including images from 150 normal individuals and 2500 patients, totaling 5150 preoperative and postoperative facial images. A machine learning model was trained to analyze these images, and 3D contour data were used to create a facial symmetry quantification system with HD computing. Additionally, 3D CBCT data from 200 patients before and after OGS were retrospectively reviewed for clinical application. The developed facial symmetry algorithm demonstrated an overall accuracy of 84.1 %. Postoperative facial symmetry scores improved significantly, with a mean score increase of 53 %, from 2.40 to 3.63. The study culminated in the creation of a web-based system that leverages HD computing and 3D contour mapping to automate facial symmetry assessment. This system offers a user-friendly interface for rapid and accurate evaluations, facilitating better communication between clinicians and patients.

Geo-spatial assessment of geomorphic characteristics of Swat Valley, Pakistan

Morphometric analysis is essential for understanding drainage networks and landform evolution, particularly in complex mountainous regions like the Swat Valley. Accurate characterization of these features aids in water resource management and mitigating climate-induced challenges. The study employed GIS and remote sensing technologies, utilizing SRTM-30m DEM data to perform a comprehensive morphometric and hypsometric analysis. Key morphometric parameters such as basin size, stream order, bifurcation ratio, drainage density, circularity ratio, and elongation ratio were calculated by using Whitebox tool. Additionally, three-dimensional parameters including basin relief, roughness number, ruggedness index, and slope were analyzed. Hypsometric analysis was conducted to examine area-elevation relationships, using empirical formulas and contour data. The basin was found to be irregularly shaped, with an elongation ratio of 6.97 and a circulatory ratio of 0.3091, indicating an elongated form. Flow accumulation data highlighted areas of rapid and moderate water flow. Hypsometric analysis showed that 14 sub-basins were in an erosional stage, 8 were mature, and the remainder were older and more resistant to erosion. The overall hypsometric integral (HI) of 0.406 reflects moderate elevation variation across the basin. The study provides critical insights into the geomorphometric controls of the Swat Valley's drainage system, contributing to improved water resource management strategies. Understanding morphometric and hypsometric characteristics is vital for adapting to climate-induced risks and safeguarding water resources in mountainous landscapes.

Development of deep learning-based novel auto-segmentation for the prostatic urethra on planning CT images for prostate cancer radiotherapy.

Urinary toxicities are one of the serious complications of radiotherapy for prostate cancer, and dose-volume histogram of prostatic urethra has been associated with such toxicities in previous reports. Previous research has focused on estimating the prostatic urethra, which is difficult to delineate in CT images; however, these studies, which are limited in number, mainly focused on cases undergoing brachytherapy uses low-dose-rate sources and do not involve external beam radiation therapy (EBRT). In this study, we aimed to develop a deep learning-based method of determining the position of the prostatic urethra in patients eligible for EBRT. We used contour data from 430 patients with localized prostate cancer. In all cases, a urethral catheter was placed when planning CT to identify the prostatic urethra. We used 2D and 3D U-Net segmentation models. The input images included the bladder and prostate, while the output images focused on the prostatic urethra. The 2D model determined the prostate's position based on results from both coronal and sagittal directions. Evaluation metrics included the average distance between centerlines. The average centerline distances for the 2D and 3D models were 2.07 ± 0.87mm and 2.05 ± 0.92mm, respectively. Increasing the number of cases while maintaining equivalent accuracy as we did in this study suggests the potential for high generalization performance and the feasibility of using deep learning technology for estimating the position of the prostatic urethra.

Diaphragm Tracking System That Superimposes Planned Diaphragm Contours on Kilovolt Cine Images During Multiple Breath-Hold Volumetric Modulated Arc Therapy for Abdominal Tumors.

We recently published a phantom validation of our diaphragm tracking system, DiaTrak, on an Elekta linear accelerator with an integrated cone-beam computed tomography (CBCT) unit for multiple breath-hold volumetric modulated arc therapy of abdominal tumors, where the diaphragm position was compared between digitally reconstructed radiography (DRR) and kilovolt (kV) projection streaming images by template matching. In the present report, the visual feedback of the diaphragm position was added to the reported system. DICOM-RT diaphragm contour data were additionally exported from a treatment planning system to the DiaTrak PC. Following phantom localization by registering the CBCT to the planning CT images, a projected diaphragm contour was overlaid on each DRR image, whereas another two projected diaphragm contours were superimposed on each kV projection cine image every 180 ms after shifting ±5 mm (set as breath-hold tolerance) in the craniocaudal direction during gantry rotation. It was visually confirmed that the projected diaphragm surface was observed within the two contour lines on the kV cine window. The diaphragm registration errors of the localized phantom were also calculated based on image cross-correlation between the DRR and the projection cine images every 180 ms. It was found that the mean diaphragm registration error was -0.29 mm with a standard deviation of 0.32 mm during the gantry rotation. In conclusion, a new interface for the 5 mm tolerance check was proposed to provide direct visual feedback, thereby giving a sense of assurance to the attending radiotherapy technologists. The calculated diaphragm registration errors were relatively small compared to the tolerance of 5 mm, and therefore it is considered clinically acceptable.

Reconstruction of the Chemical Gas Concentration Distribution Using Partial Convolution-Based Image Inpainting.

An interpolation method, which estimates unknown values with constrained information, is based on mathematical calculations. In this study, we addressed interpolation from an image-based perspective and expanded the use of image inpainting to estimate values at unknown points. When chemical gas is dispersed through a chemical attack or terrorism, it is possible to determine the concentration of the gas at each location by utilizing the deployed sensors. By interpolating the concentrations, we can obtain the contours of gas concentration. Accurately distinguishing the contours of a contaminated region from a map enables the optimal response to minimize damage. However, areas with an insufficient number of sensors have less accurate contours than other areas. In order to achieve more accurate contour data, an image inpainting-based method is proposed to enhance reliability by erasing and reconstructing low-accuracy areas in the contour. Partial convolution is used as the machine learning approach for image-inpainting, with the modified loss function for optimization. In order to train the model, we developed a gas diffusion simulation model and generated a gas concentration contour dataset comprising 100,000 contour images. The results of the model were compared to those of Kriging interpolation, one of the conventional spatial interpolation methods, finally demonstrating 13.21% higher accuracy. This suggests that interpolation from an image-based perspective can achieve higher accuracy than numerical interpolation on well-trained data. The proposed method was validated using gas concentration contour data from the verified gas dispersion modeling software Nuclear Biological Chemical Reporting And Modeling System (NBC_RAMS), which was developed by the Agency for Defense Development, South Korea.

A Binocular Color Line-Scanning Stereo Vision System for Heavy Rail Surface Detection and Correction Method of Motion Distortion.

Thanks to the line-scanning camera, the measurement method based on line-scanning stereo vision has high optical accuracy, data transmission efficiency, and a wide field of vision. It is more suitable for continuous operation and high-speed transmission of industrial product detection sites. However, the one-dimensional imaging characteristics of the line-scanning camera cause motion distortion during image data acquisition, which directly affects the accuracy of detection. Effectively reducing the influence of motion distortion is the primary problem to ensure detection accuracy. To obtain the two-dimensional color image and three-dimensional contour data of the heavy rail surface at the same time, a binocular color line-scanning stereo vision system is designed to collect the heavy rail surface data combined with the bright field illumination of the symmetrical linear light source. Aiming at the image motion distortion caused by system installation error and collaborative acquisition frame rate mismatch, this paper uses the checkerboard target and two-step cubature Kalman filter algorithm to solve the nonlinear parameters in the motion distortion model, estimate the real motion, and correct the image information. The experiments show that the accuracy of the data contained in the image is improved by 57.3% after correction.

Viscosity‐coupled curing simulation model for L‐shaped composites considering flow and compaction

AbstractComputational methods provide effective tools for the prediction of the curing behavior of composites, especially for large‐scale complex composite structures. The flow‐compaction analysis during curing receives insufficient attention in current studies. In this work, an integrated framework considering the heat transfer, the cross‐linking reaction, the flow and compaction, and the stress–strain module is established. A comprehensive resin flow termination criterion is proposed. L‐shaped composites with a stacking layer of [45/90/−45/0]s were fabricated, and the good agreement between the experimentally measured and the numerically predicted contour data validated the current model. Compared with Hubert's model, the prediction accuracy of the thickness and fiber volume content increases by 4.24% and 3.92%, respectively. Correlation analysis between the fiber volume ratio predicted by numerical methods and experimental results (94% for this method, −41% for Hubert's model) demonstrates the well‐capture of nonuniform fiber distribution in curved composites for the current strategy. The overall framework provides an accurate and promising tool for the prediction of the curing behavior of complex composite components.Highlights An integrated numerical methodology is proposed considering the resin flow and compaction. The viscosity‐temperature coupled feature of resin is considered in the permeability coefficients. A comprehensive resin flow termination criterion is proposed for an accurate prediction. The resin content, thickness, and spring‐in angle of L‐shaped laminates are characterized. The good correlation between the experimental and numerical results validates this method. Compared with the model proposed by Hubert, the prediction accuracy improves.

Esthetic analysis of upper lip morphology variation after the cross-arch fixed restoration of maxillary implant-supported prostheses via radiographic methods

Objective: To measure and analyze upper lip morphology variation before and after the cross-arch fixed restoration of the maxillary implant-supported prostheses using cone- beam CT (CBCT) to provide an esthetic objective reference for maxillary edentulous patients. Methods: There were 32 maxillary edentulous patients selected in the Department of Dental Implantology, The Affiliated Stomatological Hospital of Nanjing Medical University from January 2010 to December 2023. The CBCT data of patients with maxillary full-arch implant-supported fixed prostheses were retrospectively collected before implantation (T0), at the time of radio-guide insertion (T1), and after final repair (T2). The length and thickness of the upper lip were measured and analyzed at each time point, and the correlation between the upper lip contour data and facial esthetic parameters was analyzed. Results: After the final prosthodontic treatment, the length of the upper lip was significantly increased from (21.72±2.84) mm to (24.98±2.93) mm (t=-8.13, P<0.001) compared with that before implant treatment. The widths of the middle and vermilion of the upper lip (Sm-Hm/Ls-UP), were reduced from (13.24±1.41), (12.81±1.67) mm to (11.36±1.67), (10.21±1.69) mm, with significant differences (t=7.79, P<0001; t=9.37, P<0.001). The lower face height (Sn-Gn) was increased from (54.52±4.95) mm to (58.70±4.42) mm, with significant differences (t=-11.05, P<0.001). However, the nasolabial angle reduced significantly from 95.35°± 7.70°to 90.53°±7.28°(t=7.68, P<0.001). The width of the middle of the upper lip with a radiation guide was (10.94±1.24) mm, and it increased significantly compared with that after the final prosthesis treatment (t=-0.76, P<0.05). The proportion of straight upper lip profiles accounted for 59% (19/32), and the proportion of concave upper lip profiles accounted for 41% (13/32) after the final prosthesis treatment. In addition, the results of correlation analysis showed that the nasolabial angle variation was weak and negatively correlated with residual bone height (r=-0.37, P=0.490). Conclusions: After the treatment of the maxillary full-arch implant-supported fixed prosthesis, the length of the upper lip increases, and the thickness of the upper lip becomes significantly thin. The maxillary full-arch implant-supported fixed prostheses support upper lips to improve the patients' side appearances.

Accuracy improvement of a multi-ring beam structured inner surface measurement: via novel calibration methodology and light source optimization

Parts with a large depth-to-diameter ratio play a critical role in the military, aerospace, and automotive industries. However, accurately measuring their inner surface profile remains challenging owing to the lack of adequate and accurate sensors. We developed a multi-ring structured light system to obtain three-dimensional data of inner contours, such as inner diameters, which are crucial for ensuring component performance and safety. In this study, we proposed three simple yet effective techniques to improve the multi-ring beam structured measurement system. First, we designed a distortion correction method to calibrate the imaging system. Second, a two-step calibration approach was used to calibrate the multi-ring projection. Meanwhile, we benchmarked the effects of different light sources on image speckles. The calibration results demonstrated that the coefficient of determination (R-2) used for line fitting exceeded 0.999. Moreover, the measurement experimental results show that the uncertainty of less than 10 µm and the smallest measurable pipe inner diameter can reach 15 mm, demonstrating that our methods are promising for improving the accuracy of structured light optical sensing systems. This system satisfies the measurement requirements and can be immediately utilized to meet the high demand for inner contour measurements in industrial applications.

Assessing Spatial Representativeness of Global Flux Tower Eddy-Covariance Measurements Using Data from FLUXNET2015

Large datasets of carbon dioxide, energy, and water fluxes were measured with the eddy-covariance (EC) technique, such as FLUXNET2015. These datasets are widely used to validate remote-sensing products and benchmark models. One of the major challenges in utilizing EC-flux data is determining the spatial extent to which measurements taken at individual EC towers reflect model-grid or remote sensing pixels. To minimize the potential biases caused by the footprint-to-target area mismatch, it is important to use flux datasets with awareness of the footprint. This study analyze the spatial representativeness of global EC measurements based on the open-source FLUXNET2015 data, using the published flux footprint model (SAFE-f). The calculated annual cumulative footprint climatology (ACFC) was overlaid on land cover and vegetation index maps to create a spatial representativeness dataset of global flux towers. The dataset includes the following components: (1) the ACFC contour (ACFCC) data and areas representing 50%, 60%, 70%, and 80% ACFCC of each site, (2) the proportion of each land cover type weighted by the 80% ACFC (ACFCW), (3) the semivariogram calculated using Normalized Difference Vegetation Index (NDVI) considering the 80% ACFCW, and (4) the sensor location bias (SLB) between the 80% ACFCW and designated areas (e.g. 80% ACFCC and window sizes) proxied by NDVI. Finally, we conducted a comprehensive evaluation of the representativeness of each site from three aspects: (1) the underlying surface cover, (2) the semivariogram, and (3) the SLB between 80% ACFCW and 80% ACFCC, and categorized them into 3 levels. The goal of creating this dataset is to provide data quality guidance for international researchers to effectively utilize the FLUXNET2015 dataset in the future.

Evaluation of Recreational Facilities in Guandu section of Baoxiang River Basin Under the Background of Big Data: Based on POI

As the most important place for residents' daily leisure, the planning, layout and service level of urban recreational facilities are closely related to the quality of urban life. Taking national leisure as the benchmark and big data as the background, through the integration of POI data, vector building contour data and OSM road system data, combined with field survey, Tyson polygon is introduced to partition the urban recreational facilities in Guandu section of Baoxiang River Basin, build the urban recreational facilities system, and carry out the evaluation dimension from three aspects: quality, accessibility and coverage of urban recreational facilities, According to the main problems existing in this area, optimization suggestions are put forward, including improving the type of recreational facilities, adding the number of recreational facilities, paying attention to the fairness of recreational facilities and so on. This study has some limitations in the research scope, service scope and service population, but in view of the complex construction of the research scope, rich cultural types and certain representativeness, it can provide some reference for the evaluation and optimization of recreational facilities in similar regions.

High-speed laser centerline extraction and heterogeneous measurement architectures for low-cost hardware

Due to its real-time efficiency, stability, and reliability, structured light 3D measurement is widely used in measuring surface quality and manufacturing accuracy of industrial products. With the structured light measurement system's slow acquisition speed and low efficiency, it is difficult to meet the needs of online real-time detection, resulting in sparse 3D contour data and loss of details. To improve the algorithm's running speed and balance hardware resource consumption, a structured light extraction algorithm adapted to hardware acceleration is proposed, which executes the three low-coupling computational modules at eight-pixel parallel speed and in a pipelined manner. In order to further reduce the overall computational efficiency and deployment cost of high-speed structured light measurement tasks, an FPGA camera architecture with parallel acquisition and processing is designed, and the FPGA camera is cooperated with the CPU to form a more efficient heterogeneous measurement system. For samples of different materials and shapes, the standard deviation of sub-pixel coordinates extracted is less than 0.65. With this architecture, it is possible to process video stream with one million pixels at 500FPS, and the power consumption is 2.566w. Experimental results of measuring blocks and steel pipes show that the measurement error is below 0.15 mm. The proposed algorithm and heterogeneous processing measurement architecture are applicable in engineering, especially for high-speed real-time 3D measurement.

Interobserver delineation variability of computed tomography-based radiomic features of the parotid gland.

To assess the interobserver delineation variability of radiomic features of the parotid gland from computed tomography (CT) images and evaluate the correlation of these features for head and neck cancer (HNC) radiotherapy patients. Contrast-enhanced CT images of 20 HNC patients were utilized. The parotid glands were delineated by treating radiation oncologists (ROs), a selected RO and AccuContour auto-segmentation software. Dice similarity coefficients (DSCs) between each pair of observers were calculated. A total of 107 radiomic features were extracted, whose robustness to interobserver delineation was assessed using the intraclass correlation coefficient (ICC). Pearson correlation coefficients (r) were calculated to determine the relationship between the features. The influence of excluding unrobust features from normal tissue complication probability (NTCP) modeling was investigated for severe oral mucositis (grade ≥3). The average DSC was 0.84 (95% confidence interval, 0.83-0.86). Most of the shape features demonstrated robustness (ICC ≥0.75), while the first-order and texture features were influenced by delineation variability. Among the three observers investigated, 42 features were sufficiently robust, out of which 36 features exhibited weak correlation (|r|<0.8). No significant difference in the robustness level was found when comparing manual segmentation by a single RO or automated segmentation with the actual clinical contour data made by treating ROs. Excluding unrobust features from the NTCP model for severe oral mucositis did not deteriorate the model performance. Interobserver delineation variability had substantial impact on radiomic features of the parotid gland. Both manual and automated segmentation methods contributed similarly to this variation.

Single Person Identification and Activity Estimation in a Room from Waist-Level Contours Captured by 2D Light Detection and Ranging.

To develop socially assistive robots for monitoring older adults at home, a sensor is required to identify residents and capture activities within the room without violating privacy. We focused on 2D Light Detection and Ranging (2D-LIDAR) capable of robustly measuring human contours in a room. While horizontal 2D contour data can provide human location, identifying humans and activities from these contours is challenging. To address this issue, we developed novel methods using deep learning techniques. This paper proposes methods for person identification and activity estimation in a room using contour point clouds captured by a single 2D-LIDAR at hip height. In this approach, human contours were extracted from 2D-LIDAR data using density-based spatial clustering of applications with noise. Subsequently, the person and activity within a 10-s interval were estimated employing deep learning techniques. Two deep learning models, namely Long Short-Term Memory (LSTM) and image classification (VGG16), were compared. In the experiment, a total of 120 min of walking data and 100 min of additional activities (door opening, sitting, and standing) were collected from four participants. The LSTM-based and VGG16-based methods achieved accuracies of 65.3% and 89.7%, respectively, for person identification among the four individuals. Furthermore, these methods demonstrated accuracies of 94.2% and 97.9%, respectively, for the estimation of the four activities. Despite the 2D-LIDAR point clouds at hip height containing small features related to gait, the results indicate that the VGG16-based method has the capability to identify individuals and accurately estimate their activities.

Three-component contour dynamics model to simulate and analyze amoeboid cell motility in two dimensions.

Amoeboid cell motility is relevant in a wide variety of biomedical processes such as wound healing, cancer metastasis, and embryonic morphogenesis. It is characterized by pronounced changes of the cell shape associated with expansions and retractions of the cell membrane, which result in a crawling kind of locomotion. Despite existing computational models of amoeboid motion, the inference of expansion and retraction components of individual cells, the corresponding classification of cells, and the a priori specification of the parameter regime to achieve a specific motility behavior remain challenging open problems. We propose a novel model of the spatio-temporal evolution of two-dimensional cell contours comprising three biophysiologically motivated components: a stochastic term accounting for membrane protrusions and two deterministic terms accounting for membrane retractions by regularizing the shape and area of the contour. Mathematically, these correspond to the intensity of a self-exciting Poisson point process, the area-preserving curve-shortening flow, and an area adjustment flow. The model is used to generate contour data for a variety of qualitatively different, e.g., polarized and non-polarized, cell tracks that visually resemble experimental data very closely. In application to experimental cell tracks, we inferred the protrusion component and examined its correlation to common biomarkers: the F-actin density close to the membrane and its local motion. Due to the low model complexity, parameter estimation is fast, straightforward, and offers a simple way to classify contour dynamics based on two locomotion types: the amoeboid and a so-called fan-shaped type. For both types, we use cell tracks segmented from fluorescence imaging data of the model organism Dictyostelium discoideum. An implementation of the model is provided within the open-source software package AmoePy, a Python-based toolbox for analyzing and simulating amoeboid cell motility.

Data-driven models for the prediction of coronary atherosclerotic plaque progression/regression

Coronary artery disease is defined by the existence of atherosclerotic plaque on the arterial wall, which can cause blood flow impairment, or plaque rupture, and ultimately lead to myocardial ischemia. Intravascular ultrasound (IVUS) imaging can provide a detailed characterization of lumen and vessel features, and so plaque burden, in coronary vessels. Prediction of the regions in a vascular segment where plaque burden can either increase (progression) or decrease (regression) following a certain therapy, has remained an elusive major milestone in cardiology. Studies like IBIS-4 showed an association between plaque burden regression and high-intensity rosuvastatin therapy over 13 months. Nevertheless, it has not been possible to predict if a patient would respond in a favorable/adverse fashion to such a treatment. This work aims to (i) Develop a framework that processes lumen and vessel cross-sectional contours and extracts geometric descriptors from baseline and follow-up IVUS pullbacks; and to (ii) Develop, train, and validate a machine learning model based on baseline/follow-up IVUS datasets that predicts future percent of atheroma volume changes in coronary vascular segments using only baseline information, i.e. geometric features and clinical data. This is a post hoc analysis, revisiting the IBIS-4 study. We employed 140 arteries, from 81 patients, for which expert delineation of lumen and vessel contours were available at baseline and 13-month follow-up. Contour data from baseline and follow-up pullbacks were co-registered and then processed to extract several frame-wise features, e.g. areas, plaque burden, eccentricity, etc. Each pullback was divided into regions of interest (ROIs), following different criteria. Frame-wise features were condensed into region-wise markers using tools from statistics, signal processing, and information theory. Finally, a stratified 5-fold cross-validation strategy (20 repetitions) was used to train/validate an XGBoost regression models. A feature selection method before the model training was also applied. When the models were trained/validated on ROI defined by the difference between follow-up and baseline plaque burden, the average accuracy and Mathews correlation coefficient were 0.70 and 0.41 respectively. Using a ROI partition criterion based only on the baseline’s plaque burden resulted in averages of 0.60 accuracy and 0.23 Mathews correlation coefficient. An XGBoost model was capable of predicting plaque progression/regression changes in coronary vascular segments of patients treated with rosuvastatin therapy in 13 months. The proposed method, first of its kind, successfully managed to address the problem of stratification of patients at risk of coronary plaque progression, using IVUS images and standard patient clinical data.

Unsupervised corneal contour extraction algorithm with shared model for dynamic deformation videos: improving accuracy and noise resistance

BackgroundIn this study, an automatic corneal contour extraction algorithm with a shared model is developed to extract contours from dynamic corneal videos containing noise, which improves the accuracy of corneal biomechanical evaluation and clinical diagnoses. The algorithm does not require manual labeling and completes the unsupervised semantic segmentation of each frame in corneal dynamic deformation videos based on a fully convolutional deep-learning network using corneal geometry and texture information.ResultsWe included 1027 corneal videos at Tianjin Eye Hospital (Nankai University Affiliated Eye Hospital) from May 2020 to November 2021. The videos were obtained by the ultra-high-speed Scheimpflug camera, and then we used the shared model mechanism to accelerate the segmentation of corneal regions in videos, effectively resist noise, determine corneal regions based on shape factors, and finally achieve automatic and accurate extraction of corneal region contours. The Intersection over Union (IoU) of the extracted and real corneal contours using this algorithm reached 95%, and the average overlap error was 0.05, implying that the extracted corneal contour overlapped almost completely with the real contour.ConclusionsCompared to other algorithms, the method introduced in this study does not require manual annotation of corneal contour data in advance and can still extract accurate corneal contours from noisy corneal videos with good repeatability.

Research on trimming path for forked carrots using contour‐based machine learning methods

AbstractForked carrots are often used as animal feed or discarded directly, which affects ​the economic returns of growers and leads to resource waste and environmental pollution. After trimming, forked carrots become easier to peel and can be further processed. However, the lack of relevant studies and equipment hinders the full use of forked carrots, and the identification of fork points and determination of the trimming path are the main challenges in trimming forked carrots with unique and diverse shapes. Therefore, an automatic carrot‐trimming path recognition solution based on contour analysis and machine learning was proposed in this study to address the above challenges. Specifically, a cascaded model and a parallel model consisting of Multilayer Perceptron (MLP) and Support Vector Machine (SVM) were constructed to identify fork points, and three trimming path determination methods based on fork points and carrot contours were proposed. The results demonstrated cascaded and parallel models achieved 100% and 92.7% recall rates, respectively, with accuracy rates of 90.4% and 100% and repetition rates of 97.1% and 96.4%. Among the trimming path determination methods, both the dynamic convex hull method and the static convex hull method achieved a convexity of 94.7%, surpassing 93.1% for the slope method. The static convex hull method exhibited the fastest speed in determining the trimming path, taking only 0.0032 s per carat. The parallel model and the static convex hull method could be effectively used for online determination of the trimming path for forked carrots.Practical applicationsTrimming forked carrots enhances usability, reduces resource wastage, and mitigates environmental pollution. Leveraging contour‐based machine learning algorithms, we achieved precise fork point recognition with broad applicability. Using fork point and carrot contour data, we determined trimming paths that render carrots convex for mechanical peeling. This approach contributes to advancing sustainable agriculture by optimizing resource utilization.

Standardizing and improving dose predictions for head and neck cancers using complete sets of OAR contours.

Radiotherapy dose predictions have been trained with data from previously treated patients of similar sites and prescriptions. However, clinical datasets are often inconsistent and do not contain the same number of organ at risk (OAR) structures. The effects of missing contour data in deep learning-based dose prediction models have not been studied. The purpose of this study was to investigate the impacts of incomplete contour sets in the context of deep learning-based radiotherapy dose prediction models trained with clinical datasets and to introduce a novel data substitution method that utilizes automated contours for undefined structures. We trained Standard U-Nets and Cascade U-Nets to predict the volumetric dose distributions of patients with head and neck cancers (HNC) using three input variations to evaluate the effects of missing contours, as well as a novel data substitution method. Each architecture was trained with the original contour (OC) inputs, which included missing information, hybrid contour (HC) inputs, where automated OAR contours generated in software were substituted for missing contour data, and automated contour (AC) inputs containing only automated OAR contours. 120 HNC treatments were used for model training, 30 were used for validation and tuning, and 44 were used for evaluation and testing. Model performance and accuracy were evaluated with global whole body dose agreement, PTV coverage accuracy, and OAR dose agreement. The differences in these values between dataset variations were used to determine the effects of missing data and automated contour substitutions. Automated contours used as substitutions for missing data were found to improve dose prediction accuracy in the Standard U-Net and Cascade U-Net, with a statistically significant difference in some global metrics and/or OAR metrics. For both models, PTV coverage between input variations was unaffected by the substitution technique. Automated contours in HC and AC datasets improved mean dose accuracy for some OAR contours, including the mandible and brainstem, with a greater improvement seen with HC datasets. Global dose metrics, including mean absolute error, mean error, and percent error were different for the Standard U-Net but not for the Cascade U-Net. Automated contours used as a substitution for contour data improved prediction accuracy for some but not all dose prediction metrics. Compared to the Standard U-Net models, the Cascade U-Net achieved greater precision.

Phytoplankton Image Segmentation and Annotation Method Based on Microscopic Fluorescence.

Microscopic phytoplankton segmentation is an important part of water quality assessment. The segmentation of microscopic phytoplankton still faces challenges for computer vision, such as being affected by background impurities and requiring a large number of manual annotation. In this paper, the characteristics of phytoplankton emitting fluorescence under excitation light were utilized to segment and annotate phytoplankton contours by fusing fluorescence images and bright field images. Morphological operations were used to process microscopic fluorescence images to obtain the initial contours of phytoplankton. Then, microscopic bright field images were processed by Active Contour to fine tune the contours. Seven algae species were selected as the experimental objects. Compared with manually labeling the contour in LabelMe, the recall, precision, FI score and IOU of the proposed segmentation method are 85.3%, 84.5%, 84.7%, and 74.6%, respectively. Mask-RCNN was used to verify the correctness of labels annotated by the proposed method. The average recall, precision, F1 score and IOU are 97.0%, 86.5%, 91.1%, and 84.2%, respectively, when the Mask-RCNN is trained with the proposed automatic labeling method. And the results corresponding to manual labeling are 95.3%, 86.1%, 90.3%, and 82.8% respectively. The experimental results show that the proposed method can segment the phytoplankton microscopic image accurately, and the automatically annotated contour data has the same effect as the manually annotated contour data in Mask-RCNN, which greatly reduces the manual annotation workload.