Search Results (104)

Purpose: To develop a fully automated 2D nnU-Net pipeline for multi-class skeletal muscle segmentation (psoas, paraspinal, and abdominal wall) at the third lumbar (L3) vertebral level, and to quantitatively evaluate its diagnostic performance and reliability compared to manual segmentation. Materials and Methods: A 2D nnU-Net was trained on 164 axial L3 CT slices from the multi-institutional AMOS22 dataset, spanning diverse abdominal pathologies and multivendor imaging. To assess generalizability under severe anatomical distortion, independent external validation was performed in 50 consecutive patients with advanced liver disease from a single institution (January–December 2025; mean age, 63 ± 15 years; 32 women, 18 men), of whom 88% had moderate-to-severe ascites. Model stability was examined by comparing a five-fold ensemble with the best-performing single-fold model. Intra-observer reliability of the manual reference standard was evaluated in a random subset of 30 cases. Inter-observer agreement was additionally assessed using an independent second reader. Performance metrics included the Dice Similarity Coefficient (DSC), Pearson correlation coefficient (r), and Bland–Altman analysis for cross-sectional areas and mean attenuation. The inference workflow was deployed via a custom Streamlit-based graphical user interface (GUI). Results: In this anatomically complex external validation cohort, the 5-fold ensemble 2D nnU-Net achieved an overall mean DSC of 0.937 ± 0.043 (95% CI, 0.925–0.950), with 80% of cases achieving a mean DSC ≥ 0.90. While the mean DSC was statistically comparable to the best single-fold model (0.937, [95% CI, 0.921–0.952], p = 0.736), the ensemble strategy increased the minimum observed DSC (worst-case performance) from 0.720 to 0.822. Class-specific external validation performance for the 5-fold ensemble was highest for the paraspinal muscles (DSC: 0.960; 95% CI, 0.952–0.967), followed by the psoas muscles (DSC: 0.941; 95% CI, 0.927–0.956), and lowest for the anatomically complex abdominal wall muscles (DSC: 0.911; 95% CI, 0.893–0.929). Comparison between the ensemble model and manual segmentation yielded a Pearson correlation of r = 0.955 (p < 0.001) for total skeletal muscle area, with a mean bias of +7.17 cm². Intra- and inter-observer agreements for the manual reference standard demonstrated correlation coefficients of r = 0.995 and 0.090 for total areas, respectively. The automated pipeline required 3–5 s per case for inference and quantitative reporting, compared to 3–5 min for manual segmentation. Conclusions: In patients with advanced liver disease and substantial anatomical distortion from ascites, an ensemble-based 2D nnU-Net provides high quantitative agreement with manual L3 skeletal muscle segmentation, while mitigating lower-bound (worst-case) errors relative to single-fold models. Integration with a dedicated GUI enables substantial time savings and supports scalable quantitative body composition measurement. Full article

Background and Objectives: Acetabular fracture surgery is associated with substantial perioperative blood loss and prolonged operative time. Routine preoperative pelvic computed tomography (CT) carries information about body composition that is not currently exploited for risk stratification. We tested whether (i) CT-defined pelvic sarcopenia is associated with lower preoperative haemoglobin and (ii) preoperative subcutaneous fat cross-sectional area (CSA) is independently associated with operative time, after adjustment for surgical approach, age, fracture complexity and sarcopenia status. Materials and Methods: In this single-centre retrospective cohort study, 48 adults (37 men, 11 women; mean age 40.2 ± 16.5 years) who underwent open reduction and internal fixation (ORIF) for unilateral acetabular fractures between 2016 and 2024 were included. Pelvic muscle and subcutaneous fat CSAs were measured on the contralateral side of preoperative CT images using ImageJ. Sarcopenia was defined as an internal, cohort-relative classification based on the sex-specific bottom tertile of psoas CSA. Normality was assessed by Shapiro–Wilk testing; Pearson or Spearman correlation was used accordingly, and the 36 pairwise correlations were controlled with the Benjamini–Hochberg false-discovery-rate procedure. The multivariable model used ordinary least squares regression with heteroscedasticity-consistent (HC3) standard errors and a median quantile-regression robustness check. Results: Sarcopenic patients (n = 17) had significantly lower preoperative haemoglobin (12.63 ± 1.24 vs. 14.00 ± 1.53 g/dL; p = 0.002; Cohen’s d = 0.96). The absolute perioperative haemoglobin drop was numerically smaller in the sarcopenic group (ΔHb 1.64 ± 0.91 vs. 2.46 ± 1.87 g/dL) but did not reach statistical significance (p = 0.079); estimated blood loss (p = 0.258) and transfusion requirement (p = 0.567) did not differ between groups. Pelvic muscle CSAs correlated positively with preoperative haemoglobin (all q < 0.05 after Benjamini–Hochberg correction). In the multivariable model (F[6, 41] = 3.71, p = 0.005; adjusted R² = 0.26; all variance inflation factors 1.06–1.26), subcutaneous fat CSA (B = +0.25 min/cm², p = 0.004) and the modified Stoppa approach (vs. Kocher–Langenbeck; +65 min, p = 0.001) were independently associated with operative time. Conclusions: In this exploratory retrospective cohort, routine preoperative pelvic CT contained two body-composition signals that may warrant prospective evaluation: pelvic sarcopenia, which was associated with lower baseline haemoglobin, and subcutaneous adiposity, which was associated with longer operative time in the primary regression model. Both signals require confirmation—the sarcopenia–bleeding relationship was not statistically significant, and the subcutaneous fat association was attenuated under robust inference. These findings are hypothesis-generating; prospective multicentre validation with height-normalised sarcopenia thresholds and body mass index is required before clinical implementation. Full article

Background: Early postoperative changes in creatinine-based estimated glomerular filtration rate (eGFR) after bariatric surgery can be misread as a kidney injury. During rapid weight loss, indexing eGFR to a fixed body surface area (BSA) of 1.73 m² may alter apparent trajectories. We compared absolute (mL/min) and BSA-indexed (mL/min/1.73 m²) eGFR changes after sleeve gastrectomy, stratified by baseline glomerular hyperfiltration (GH). Methods: In this retrospective cohort of 145 adults undergoing laparoscopic sleeve gastrectomy, serum creatinine was obtained at baseline (≤30 days pre-op) and 3 months (post-op days 75–105). Indexed eGFR was calculated with the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) 2021 creatinine equation; BSA with the Mosteller formula; and absolute eGFR as indexed eGFR × (BSA/1.73). GH was defined as indexed eGFR ≥ 120 mL/min/1.73 m². A REML mixed-effects model (Group, Time, Group × Time) with patient-cluster bootstrap inference was used. An age-adjusted sensitivity model including Age and Age × Time was also fitted. Results: Fifty-four participants (37%) met the GH criteria. Absolute eGFR declined by −26.6 mL/min in GH versus −17.3 mL/min in non-GH (difference-in-differences [DiD] −9.3 mL/min; 95% CI −13.9 to −4.7; p < 0.001). The indexed eGFR changes were smaller (−4.2 vs. −0.5 mL/min/1.73 m²; DiD −3.7; 95% CI −7.3 to −0.03; p = 0.048; bootstrap p_sign = 0.052). In the age-adjusted sensitivity model, the Group × Time interaction for absolute eGFR attenuated but remained statistically significant (−6.57 mL/min; 95% CI, −13.09 to −0.06; p = 0.048), whereas the corresponding interaction for indexed eGFR was attenuated and no longer statistically significant (−3.99 mL/min/1.73 m²; 95% CI −9.15 to 1.16; p = 0.129). Conclusions: Within three months after sleeve gastrectomy, participants with higher baseline indexed filtration showed a larger decline in absolute eGFR but only a small change in indexed eGFR. These results show that early postoperative creatinine-based eGFR trajectories are scale dependent and should be interpreted cautiously during rapid weight loss. Because postoperative acute kidney injury (AKI) was not adjudicated and direct kidney function markers were unavailable, this study does not distinguish physiological hemodynamic change from structural kidney injury. Reporting both absolute and indexed eGFR may improve early postoperative interpretation and help align dosing decisions with rapid changes in body size. Full article

Objective: Back pain is a multifactorial condition commonly associated with degenerative spinal changes. Spondylophytes are frequent outgrowths of the vertebral bodies that may be influenced by arterial hypertension via a possible increased pulsation of the aorta and its effects on bone remodeling. If it can be demonstrated that an increased pulse pressure in the aorta due to hypertension promotes the growth of spondylophytes and thereby increases the likelihood of back pain, future studies may investigate how the effectiveness of blood pressure management can be improved in order to reduce the prevalence of degenerative changes in the spine and, consequently, prevent back pain. This study investigated the association between arterial hypertension and thoracic spondylophyte formation using whole-body MRI data from the population-based Study of Health in Pomerania (SHIP). Materials and Methods: Spondylophyte presence and area were assessed for their association with hypertension status in 859 SHIP-START-3 participants who underwent whole-body MRI. Right-sided spondylophytes at T8-T11 were measured on axial T2-weighted sequences. Hypertension was defined by self-report or antihypertensive medication use; a sensitivity analysis was conducted using the 2024 European Society of Cardiology definition (systolic blood pressure ≥ 140 mmHg). Multivariate regression models adjusted for age, sex, obesity, and smoking were used to assess associations. Machine learning algorithms were applied for validation. Results: Spondylophytes were present in 87.7% of participants. Hypertension was significantly associated with spondylophyte presence (OR = 2.07, 95% CI: 1.15–3.81) but not consistently associated with spondylophyte size. Spondylophyte size increased from T8 to T11, and was associated with age, male sex, and obesity. Sensitivity analyses widely confirmed robustness of the analysis. Conclusions: This population-based MRI study investigates the still insufficiently studied relationship between arterial hypertension and the formation of thoracic spondylophytes. The findings are consistent with the hypothesis that hypertension may be associated with spinal bone remodelling, though causal inference remains limited by the cross-sectional study design. Further longitudinal studies are needed to clarify causality and clinical relevance for spinal degeneration and back pain. Full article

Background: Normal-weight obesity (NWO) is a nutritional status in which individuals have a normal body mass index (BMI) with a high percentage of body fat (%BF). However, the impact of elevated %BF on cardiometabolic risk remains unclear. This study aimed to evaluate whether NWO is associated with worse cardiometabolic risk markers and scores. Methods: We conducted a cross-sectional study using a convenience sample of employees from a public hospital. Participants aged ≥18 years with a BMI between 18.5–24.9 kg/m² were included in the study. %BF was categorized according to sex and age (InBody720). Normal weight and normal %BF (NWNB) and NWO were defined using cutoff points. Body composition, serum biochemical and inflammatory markers, hemodynamics, and autonomic function were considered cardiometabolic risk markers. The visceral fat area (VFA), atherogenic coefficient (AC), atherogenic index of plasma (AIP), body shape index (ABSI), and Framingham Risk (FR) score were considered cardiometabolic risk scores. Statistical significance was set at p < 0.05. Results: Of the 228 eligible participants, 52 met the inclusion criteria (NWNB, N = 29 and NWO, N = 23). Participants with NWO presented worse values of lipid profiles, anthropometric measurements, hemodynamic parameters, and autonomic function indices. After adjustment for age and sex, NWO remained associated with selected cardiometabolic markers, particularly LDL-c, triglycerides, and autonomic indices, whereas body composition findings should be interpreted as confirmatory of the phenotype. Conclusions: In this cross-sectional secondary analysis, NWO was associated with worse cardiometabolic markers and selected risk scores compared with NWNB. These findings support an unfavorable cardiometabolic profile in individuals with NWO, but do not allow inferences about future cardiometabolic events or causal relationships. Longitudinal studies are needed to clarify its prognostic significance. Full article

Natural hydrogen is regarded as a potential resource for the global energy transition, and its accumulation is closely linked to water–rock reactions involving Fe²⁺ bearing minerals and effective sealing conditions. The Liaohe Basin, located on the northeastern margin of the North China Craton within a key metallogenic belt, is surrounded by sedimentary-metamorphic iron deposits and is a potential area for natural hydrogen accumulation. In this study, aeromagnetic and satellite gravity data were integrated to estimate basement depth through gravity interface inversion, followed by three-dimensional magnetic susceptibility and density inversion and structural–mineralization correlation analysis. The results reveal strong basement heterogeneity. Iron-rich anomalous bodies show clustered and belt-like to dome-like distributions, mainly along the transitional zone between deep depressions and basement uplifts. Combined density–magnetic zonation suggests that high-density, high-magnetic units may correspond to iron-rich bodies, whereas high-magnetic, low-density units likely indicate fractured and altered fluid pathways. Based on the measured results of surface hydrogen concentration, it is inferred that the high magnetic anomaly in the uplift transition zone at the edge of the depression might be the coupling area of iron-rich rock bodies and channel zones, which is the priority response area of natural hydrogen in the Liaohe Basin, China. Full article

The dissemination of sustainable knowledge within the domain of higher education has grown exponentially since the implementation of the UN’s SDGs; however, the body of evidence is currently fragmented across various institutional and educational sectors. This research synthesizes review-level evidence on how institutions of higher education provide for the dissemination of sustainable knowledge and develop the competencies necessary to support it, through a narrative literature review with a supporting structured Web of Science search, transparent narrowing, and interpretive thematic synthesis. An evidence set of focused relevance (2015–2025) was established from an initial total of 6604 records and through the subsequent full-text analysis yielded a final corpus of 63 review articles. Two dominant theme categories were identified: (i) Institutional Embedding and Governance Level Integration and (ii) Educational Level Implementation. A third area of investigation mapped the development of the discipline through both bibliometric and narrative reviews. A common cross-cutting constraint is that specific links between mechanisms/outcomes, as well as comparative analyses of student outcome metrics across studies, are not uniformly documented, which limits cumulative inferences about effective practices. Thus, greater clarity is needed regarding linkages between competence objectives, curriculum, pedagogy, assessment, and measurable outcomes. Additionally, the governance conditions are frequently referenced as enabling factors. Full article

Accurate lithologic maps derived from satellite imagery underpin structural interpretation, mineral exploration, and geohazard assessment. However, automated mapping in complex terranes remains challenging because spectrally similar units, narrow anisotropic bodies, and ambiguous contacts can degrade boundary fidelity. In this study, we propose SegNeXt-HFCA, a hierarchical multiscale fusion network with coordinate attention for lithologic segmentation from a Sentinel-2/DEM feature stack. The model builds on SegNeXt and introduces a hierarchical multiscale encoder with coordinate attention to jointly capture fine textures and scene-level structure. It further adopts a class-frequency-aware hybrid loss that combines boundary-weighted online hard-example mining cross-entropy with Lovász-Softmax to better handle long-tailed classes and ambiguous contacts. In addition, we employ a robust training and inference scheme, including entropy-guided patch sampling, exponential moving average of parameters, test-time augmentation, and a DenseCRF-based post-refinement. Two study areas in the Beishan orogen, northwestern China (Huitongshan and Xingxingxia), are used to evaluate the method with a unified 10-channel Sentinel-2/DEM feature stack. Compared with U-NetFormer, PSPNet, DeepLabV3+, DANet, LGMSFNet, SegFormer, BiSeNetV2, and the SegNeXt backbone, SegNeXt-HFCA improves mean intersection-over-union (mIoU) by about 3.8% in Huitongshan and 2.6% in Xingxingxia, respectively, and increases mean pixel accuracy by approximately 3–4%. Qualitative analyses show that the proposed framework better preserves thin-unit continuity, clarifies lithologic contacts, and reduces salt-and-pepper noise, yielding geologically more plausible maps. These results demonstrate that hierarchical multiscale fusion with coordinate attention, together with class- and boundary-aware optimization, provides a practical route to robust lithologic mapping in structurally complex regions. Full article

The Internet of Military Things (IoMT) relies on Long-Range Wide Area Networks (LoRaWAN) for low-power, long-range communication in critical applications like border security and soldier health monitoring. However, conventional priority-based flow control mechanisms, which rely on static classification thresholds, lack the adaptability to handle the nuanced, continuous nature of physiological data and dynamic network states. To overcome this rigidity, this paper introduces a novel, domain-adaptive Fuzzy Logic Flow Control (FFC) protocol specifically tailored for LoRaWAN-based IoMT. While employing established Mamdani inference, the FFC system innovatively fuses multi-parameter physiological data (body temperature, blood pressure, oxygen saturation, and heart rate) into a continuous Health Score, which is then mapped via a context-optimised sigmoid function to dynamic transmission intervals. This represents a novel application-layer semantic integration with LoRaWAN’s constrained MAC and PHY layers, enabling cross-layer flow optimisation without protocol modification. Simulation results confirm that FFC significantly enhances reliability and energy efficiency while reducing latency relative to traditional static priority architectures. Seamlessly integrated into the NS-3 LoRaWAN simulation framework, the FFC protocol demonstrates superior performance in IoMT communications. Simulation results confirm that FFC significantly enhances reliability and energy efficiency while reducing latency compared with traditional static priority-based architectures. It achieves this by prioritising high-priority health telemetry, proactively mitigating network congestion, and optimising energy utilisation, thereby offering a robust solution for emergent, health-critical scenarios in resource-constrained environments. Full article

Background/Objectives: Recessive dystrophic epidermolysis bullosa (RDEB) is a severe congenital genodermatosis characterized by skin and mucosa fragility, chronic inflammation, recurrent infections and high nutritional demands due to increased metabolism and epithelial barrier-related losses, placing patients at risk of zinc deficiency. We aimed to investigate the clinical relevance and biochemical determinants of zinc deficiency as a potentially modifiable contributor to disease burden in RDEB. Methods: In this cross-sectional study (n = 84), serum zinc levels were analyzed in association with sex, age, disease severity, percentage of body surface area (BSA) affected, inflammatory markers, infection burden, and common clinical complications including anemia and growth impairment. Results: Zinc deficiency, defined as levels below 670 µg/L, was identified in 35% of patients and became more frequent after age 5 and during adulthood, particularly among those with more severe disease. Deficiency was strongly associated with anemia, inflammation, infection burden, growth impairment, and extensive skin involvement. A revised cutoff of 780 µg/L is proposed, showing improved diagnostic performance for identifying patients at risk of systemic complications, and offering a more suitable threshold for starting preventive supplementation. Multivariate logistic modeling confirmed that low serum zinc independently predicted anemia risk, alongside transferrin saturation and C- reactive protein levels. Serum albumin was identified as the strongest determinant of zinc levels, partially mediating the effects of inflammation and skin involvement. Conclusions: These findings identify serum zinc as a clinically relevant marker of nutritional status and complication burden in RDEB. While no causal or therapeutic effects can be inferred from this cross-sectional study, the strong and biologically plausible associations observed suggest a rationale for systematic monitoring and correction of zinc deficiency as part of comprehensive supportive care, and warrant prospective studies to assess clinical benefit. Full article

Objectives: This study investigated the prevalence and severity of musculoskeletal injuries within South Africa’s most popular karate style, Shotokan, a previously unexamined area. As an exploratory study, it aimed to generate hypotheses by determining the prevalence, severity, and nature of these injuries to address this significant gap in the national combat sports literature. Methods: A descriptive, cross-sectional design was employed, utilizing a confidential online questionnaire distributed through various Shotokan organizations. The study gathered 155 responses (26.85% response rate). Results: The findings revealed a high injury prevalence, with 47.3% of participants reporting at least four injuries. These injuries occurred equally in training and competition (56.5%) and developed both acutely and over time (53.4%). Experienced practitioners at the Shodan level were particularly affected. The knee was the most frequently injured body part (11.6%), and muscle strains were the most common injury type (19.3%). Notably, 26.2% of karatekas continued training despite being injured. A significant weak positive correlation was found between years of training experience and injury levels (rs = 0.275, p = 0.007). However, no significant associations were found between injury prevalence and age, BMI, or training frequency. General practitioners were the most consulted healthcare professionals (22.0%). Conclusions: This study establishes a high prevalence of musculoskeletal injuries among South African Shotokan karatekas, particularly associated with experienced practitioners. These findings are hypothesis-generating, and the cross-sectional design precludes causal inferences. The data provides a crucial foundation for future longitudinal research to investigate causality and for developing evidence-based injury prevention protocols, particularly for the knee. Full article

The phenotypic features and yield of Lentinula edodes fruiting bodies are key metrics in breeding, cultivation, and quality evaluation. To overcome the inefficiency and physical damage associated with manual measurement, this paper proposes an automated approach using a lightweight YOLOv11-Seg model. On the basis of the YOLOv11-Seg model, the ShuffleNetV2 network, the C3k2-FasterBlock feature extraction module, and the convolutional block attention module (CBAM) were introduced to construct a lightweight YOLO-SFCB model, which significantly reduced the complexity and computational cost of the model. The experimental results show that the parameters, floating point operations (FLOPs), and mAP50-95 of the YOLO-SFCB model reach 2.0 M, 7.8 G, and 80.5%, respectively, while the GPU-based inference time is only 15.7 ms. Compared with the original model, parameters and FLOPs were reduced by 29% and 25%, inference time was shortened by 9.8%, and mAP50-95 increased by 0.9%. Based on the YOLO-SFCB model, OpenCV was used to extract the minimum rotation circumscribed rectangle of the stipe and pileus segmentation areas, and the stipe height, stipe diameter, pileus width, and pileus thickness were measured; the average residual is less than 0.35 mm. Finally, using the least squares method, a yield prediction model for Lentinula edodes fruiting bodies was developed. The average prediction errors for fresh weight and dry weight were controlled within 0.5 g and 0.15 g, respectively. The YOLO-SFCB model and the method for extracting phenotypic features and predicting yield of Lentinula edodes proposed in this study can help promote the development of Lentinula edodes breeding and cultivation and stabilize market supply and demand. Full article

Yield prediction in longline aquaculture is essential for evaluating environmental impacts, facilitating risk assessment, and promoting sustainable management in fisheries. However, since cultured organisms in longline aquaculture are submerged and cannot be directly observed, existing yield prediction approaches are mostly based on indirect environmental proxies, which often lead to unsatisfactory accuracy. The Shadow Geometry Inversion for Aquaculture (SGIA) method enables direct and accurate yield estimation in longline aquaculture by utilizing the submergence state of buoys to infer load, which is determined by the weight of the cultured organisms and estimated by shadow lengths combined with solar altitude angles and buoy physical parameters in high-resolution unmanned aerial vehicle (UAV) imagery. Experiments have been conducted in a water body located in Shanghai and Fuding to validate the effectiveness of the SGIA method. The best results were achieved under solar altitudes of 10–25° and calm water conditions. Under these conditions, the SGIA-predicted yields closely matched the measured loads in the Shanghai controlled experiment (R² = 0.985, MAPE = 9.19%). In the Fuding field application, the model effectively captured spatial variations in buoy loads across the farming area, demonstrating its practicality and scalability for large-scale yield mapping in real aquaculture environments. Full article

Background: The early detection of insulin resistance (IR) is critical for the prevention of type 2 diabetes and cardiometabolic diseases. The ECORE-BF index is a simple anthropometric tool for estimating body fat percentage and overweight. However, its potential utility as a predictor of IR risk has not been previously evaluated in large populations using validated IR indices. Methods: This cross-sectional study included 418,343 Spanish workers (172,282 women and 246,061 men) who underwent occupational health evaluations. The ECORE-BF index was calculated for all participants, and its association with four validated surrogate markers of IR was analyzed: the triglyceride–glucose index (TyG), TyG-BMI, METS-IR, and SPISE. Subjects were classified into normal or high-risk IR groups based on established cut-off values. We evaluated the mean ECORE-BF values across groups, the prevalence of ECORE-BF-defined obesity, and the diagnostic performance of ECORE-BF using receiver operating characteristic (ROC) curve analysis. Results: Participants with elevated IR index values had significantly higher mean ECORE-BF scores than those with normal values (p < 0.001). The prevalence of ECORE-BF-defined obesity was substantially higher in all high-risk IR groups, exceeding 99% for METS-IR and SPISE in both sexes. ROC analysis demonstrated the high diagnostic accuracy of ECORE-BF in predicting elevated IR risk, with area under the curve (AUC) values ranging from 0.698 (TyG in men) to 0.992 (METS-IR in women). Sensitivity and specificity were also high, particularly for TyG-BMI, SPISE, and METS-IR, with optimal Youden indices above 0.75. Conclusions: ECORE-BF demonstrated high accuracy as a non-invasive tool for identifying individuals at increased insulin resistance risk; however, due to the cross-sectional design, predictive value for incident disease cannot be inferred. Its simplicity, cost-effectiveness, and high diagnostic accuracy support its potential utility in large-scale screening programs for early detection of metabolic risk. Full article

The identification of estrus at the right moment allows for a higher success of fecundity with artificial insemination. Evaluating changes in body surface temperature of sows during the estrus period using an infrared thermography camera (ITC) can provide an accurate model to predict these changes. This pilot study comprised nine crossbred Large White x Landrace sows, providing 59 data records for analysis. Observed changes in the behavior and physiological signs of the sows signaled the identification of estrus. Images of the ocular area, ear tips, breast, back, vulva, and perianal area were collected with the ITC. The images were analyzed using the FLIR Thermal Studio Starter software. Infrared mean temperatures were reported and compared using ANOVA and Tukey–Kramer tests (p < 0.05). Supervised machine learning models were tested using random forest (RF), Conditional inference trees (Ctree), Partial least squares (PLS), and K-nearest neighbors (KNN), and the method performance was measured using a confusion matrix. The orbital region showed significant differences between estrus and non-estrus states in sows. In the confusion matrix, the algorithm predicted estrus with 87% accuracy in the test set, which contained 40% of the data, when agitated behavior was combined with orbital area temperature. These findings suggest the potential for integrating behavioral and physiological observations with orbital thermography and machine learning to detect estrus in sows under field conditions accurately. Full article