Search

Human pose estimation is a critical research area in the field of computer vision, playing a significant role in applications such as human-computer interaction, behavior analysis, and action recognition. In this paper, we propose a U-shaped keypoint detection network (DAUNet) based on an improved ResNet subsampling structure and spatial grouping mechanism. This network addresses key challenges in traditional methods, such as information loss, large network redundancy, and insufficient sensitivity to low-resolution features. DAUNet is composed of three main components. First, we introduce an improved BottleNeck block that employs partial convolution and strip pooling to reduce computational load and mitigate feature loss. Second, after upsampling, the network eliminates redundant features, improving the overall efficiency. Finally, a lightweight spatial grouping attention mechanism is applied to enhance low-resolution semantic features within the feature map, allowing for better restoration of the original image size and higher accuracy. Experimental results demonstrate that DAUNet achieves superior accuracy compared to most existing keypoint detection models, with a mean PCKh@0.5 score of 91.6% on the MPII dataset and an AP of 76.1% on the COCO dataset. Moreover, real-world experiments further validate the robustness and generalizability of DAUNet for detecting human bodies in unknown environments, highlighting its potential for broader applications. [ABSTRACT FROM AUTHOR]

Background: The requirement for precise and effective delivery of the actual dose to the patient grows along with the complexity of breast cancer radiotherapy. Dosimetry during treatment has become a crucial component of guaranteeing the efficacy and security. Purpose: To propose a dosimetry method during breast cancer radiotherapy based on body surface changes. Methods: A total of 29 left breast cancer radiotherapy cases were retroactively retrieved from an earlier database for analysis. Non‐rigid image registration and dose recalculation of the planning computed tomography (CT) referring to the Cone‐beam computed tomography were performed to obtain dose changes. The study used 3D point cloud feature extraction to characterize body surface changes. Based on the correlation proof, a mapping model is developed between body surface changes and dose changes using neural network framework. The MSE metrics, the Euclidean distances of feature points and the 3D gamma pass rate metric were used to assess the prediction accuracy. Results: A strong correlation exist between body surface changes and dose changes (first canonical correlation coefficient = 0.950). For the dose deformation field and dose amplitude difference in the test set, the MSE of the predicted and actual values were 0.136 pixels and 0.229 cGy, respectively. After deforming the planning dose into a deformed one, the feature points' Euclidean distance between it and the recalculated dose changes from 9.267 ± 1.879 mm to 0.456 ± 0.374 mm. The 3D gamma pass rate of 90% or higher for the 2 mm/2% criteria were achieved by 80.8% of all cases, with a minimum pass rate of 75.9% and a maximum pass rate of 99.6%. Pass rate for the 3 mm/2% criteria ranged from 87.8% to 99.8%, with 92.3% of the cases having a pass rate of 90% or higher. Conclusions: This study provides a dosimetry method that is non‐invasive, real‐time, and requires no additional dose for breast cancer radiotherapy. [ABSTRACT FROM AUTHOR]

Background: Retrospective studies indicate that radiation damage to left anterior descending coronary artery (LAD) may be critical for late‐stage radiation‐induced cardiac morbidity. Developing a method that accurately depicts LAD motion and perform dose assessment is crucial. Purpose: To construct a generalized cardiac surface motion model for LAD dose assessment in left breast cancer radiotherapy. Methods: Cine MRI of 25 cases were divided into training and testing sets for model construction, and five external cases were gathered for generalization validation. Motion prediction from average intensity projection images (AIP) surface point cloud to that of each phase was realized by mapping the relationship between datum points and corresponding points with statistical shape modeling (SSM). Root mean square error (RMSE) for predicted corresponding points and Euclidean distance (ED) for predicted surface point cloud were used to assess model's accuracy. LAD dose assessment for 10 left breast cancer radiotherapy cases was perform by model application. Results: The RMSE in testing cases and external cases were 0.209 ± 0.020 mm to 0.841 ± 0.074 mm and 0.895 ± 0.093 mm to 1.912 ± 0.138 mm, respectively; while the ED were 1.399 ± 0.029 mm to 1.658 ± 0.100 mm, 1.571 ± 0.080 mm to 1.779 ± 0.104 mm, respectively, proving the generalized model's high accuracy. The volume of LAD characterizing motion range (WPLAD) (2.392 ± 0.639 cm3) was approximately twice that of LAD from superimposed images (SPLAD) (0.927 ± 0.326 cm3) with p < 0.05, and the former's Dmax (3582.06 ± 575.92 cGy) was significantly larger than latter's (3222.71 ± 665.37 cGy) (p < 0.05). While WPLAD's Dmean (1408.06 ± 413.06 cGy) was slightly smaller than that of SPLAD (1504.15 ± 448.03 cGy), the difference did not reach statistical significance (p > 0.05). WPLAD's V20 (23.42% ± 16.62%) was less than SPLAD's (29.18% ± 21.07%) with p < 0.05, but their comparison in V30 and V40 did not yield statistically significant results. It implies the conventional LAD dose assessment ignores motion impact and may not be justified. Conclusions: The generalized cardiac surface motion model informs LAD dose accurate assessment in left breast cancer radiotherapy. [ABSTRACT FROM AUTHOR]

Feature selection (FS) is a classic and challenging optimization task in most machine learning and data mining projects. Recently, researchers have attempted to develop more effective methods by using metaheuristic methods in FS. To increase population diversity and further improve the effectiveness of the beluga whale optimization (BWO) algorithm, in this paper, we propose a multi-strategies improved BWO (MSBWO), which incorporates improved circle mapping and dynamic opposition-based learning (ICMDOBL) population initialization as well as elite pool (EP), step-adaptive Lévy flight and spiral updating position (SLFSUP), and golden sine algorithm (Gold-SA) strategies. Among them, ICMDOBL contributes to increasing the diversity during the search process and reducing the risk of falling into local optima. The EP technique also enhances the algorithm′s ability to escape from local optima. The SLFSUP, which is distinguished from the original BWO, aims to increase the rigor and accuracy of the development of local spaces. Gold-SA is introduced to improve the quality of the solutions. The hybrid performance of MSBWO was evaluated comprehensively on IEEE CEC2005 test functions, including a qualitative analysis and comparisons with other conventional methods as well as state-of-the-art (SOTA) metaheuristic approaches that were introduced in 2024. The results demonstrate that MSBWO is superior to other algorithms in terms of accuracy and maintains a better balance between exploration and exploitation. Moreover, according to the proposed continuous MSBWO, the binary MSBWO variant (BMSBWO) and other binary optimizers obtained by the mapping function were evaluated on ten UCI datasets with a random forest (RF) classifier. Consequently, BMSBWO has proven very competitive in terms of classification precision and feature reduction. [ABSTRACT FROM AUTHOR]

Purpose: The positional accuracy of MLC is an important element in establishing the exact dosimetry in VMAT. We comprehensively analyzed factors that may affect MLC positional accuracy in VMAT, and constructed a model to predict MLC positional deviation and estimate planning delivery quality according to the VMAT plans before delivery. Methods: A total of 744 "dynalog" files for 23 VMAT plans were extracted randomly from treatment database. Multi‐correlation was used to analyzed the potential influences on MLC positional accuracy, including the spatial characteristics and temporal variability of VMAT fluence, and the mechanical wear parameters of MLC. We developed a model to forecast the accuracy of MLC moving position utilizing the random forest (RF) ensemble learning method. Spearman correlation was used to further investigate the associations between MLC positional deviation and dosage deviations as well as gamma passing rates. Results: The MLC positional deviation and effective impact factors show a strong multi‐correlation (R = 0.701, p‐value < 0.05). This leads to the development of a highly accurate prediction model with average variables explained of 95.03% and average MSE of 0.059 in the 5‐fold cross‐validation, and MSE of 0.074 for the test data was obtained. The absolute dose deviations caused by MLC positional deviation ranging from 12.948 to 210.235 cGy, while the relative volume deviation remained small at 0.470%–5.161%. The average MLC positional deviation correlated substantially with gamma passing rates (with correlation coefficient of −0.506 to −0.720 and p‐value < 0.05) but marginally with dosage deviations (with correlation coefficient < 0.498 and p‐value > 0.05). Conclusions: The RF predictive model provides a prior tool for VMAT quality assurance. [ABSTRACT FROM AUTHOR]

The development of rigid yet fatigue‐resistant elastomers possess a critical challenge, necessitating materials capable of enduring substantial mechanical loads while maintaining structural integrity through cycles of deformations. Despite soft materials achieving a fatigue threshold of ∼1000 J m−2, their modulus, typically ∼1 MPa, restricts their applications. This study presents a breakthrough in designing rigid, fatigue‐resistant elastomers by controlling nanosheet orientation within the network using a shear force strategy. The elastomers achieve a high fatigue threshold of 1900 J m−2 while maintaining a high elastic modulus of 8 MPa. These superior mechanical properties are attributed to the synergistic effects of microscale alignment of nanosheets and robust interfacial bonding between the polymer chains and nanosheets at the nanoscale. When the nanosheet volume fraction is below the percolation threshold, well‐dispersed nanosheets transfer load effectively through strong interfacial bonding with the polymer, significantly suppressing fatigue crack propagation. Conversely, above the percolation threshold, agglomerated nanosheets create defects in the matrix, reducing load‐bearing capacity. The innovative elastomers demonstrate potential for a wide range of applications, including automotive components, actuators, and stretchable electronic devices, where high mechanical stability and load‐bearing capacity are essential. [ABSTRACT FROM AUTHOR]

Background: Stroke etiology could influence the outcomes in patients with basilar-artery occlusion (BAO). This study aimed to evaluate the differences in efficacy and safety of best medical treatment (BMT) plus endovascular treatment (EVT) versus BMT alone in acute BAO across different stroke etiologies., Methods: The study was a post hoc analysis of the ATTENTION trial (Trial of Endovascular Treatment of Acute Basilar-Artery Occlusion), which was a multicenter, randomized trial at 36 centers in China from February 2021 to September 2022. Patients with acute BAO were classified into 3 groups according to stroke etiology (large-artery atherosclerosis [LAA], cardioembolism, and undetermined cause/other determined cause [UC/ODC]). The primary outcome was a favorable outcome (modified Rankin Scale score of 0-3) at 90 days. Safety outcomes included symptomatic intracranial hemorrhage and 90-day mortality., Results: A total of 340 patients with BAO were included, 150 (44.1%) had LAA, 72 (21.2%) had cardioembolism, and 118 (34.7%) had UC/ODC. For patients treated with BMT plus EVT and BMT alone, respectively, the rate of favorable outcome at 90 days was 49.1% and 23.8% in the LAA group (odds ratio, 3.08 [95% CI, 1.38-6.89]); 52.2% and 30.8% in the cardioembolism group (odds ratio, 2.45 [95% CI, 0.89-6.77]); and 37.5% and 17.4% in the UC/ODC group (odds ratio, 2.85 [95% CI, 1.16-7.01]), with P =0.89 for the stroke etiology×treatment interaction. The rate of symptomatic intracranial hemorrhage in EVT-treated patients with LAA, cardioembolism, and UC/ODC was 8.3%, 2.2%, and 3.2%, respectively, and none of the BMT-treated patients. Lower 90-day mortality was observed in patients with EVT compared with BMT alone across 3 etiology groups., Conclusions: Among patients with acute BAO, EVT compared with BMT alone might be associated with favorable outcomes and lower 90-day mortality, regardless of cardioembolism, LAA, or UC/ODC etiologies. The influence of stroke etiology on the benefit of EVT should be explored by further trials., Registration: URL: https://www.clinicaltrials.gov; Unique identifier: NCT04751708., Competing Interests: Dr Nguyen discloses advisory board with Brainomix, Aruna Bio; associate editor of Stroke. The other authors report no conflicts.

Background: Tumor hypoxia has been frequently detected in nasopharyngeal carcinoma (NPC) and is intently associated with therapeutic resistance. The aim of the study is to establish a clonogenically stable hypoxia-inducible dual reporter model and apply it to investigate the effect of tumor hypoxia on DNA double strand break (DSB) and synergistic effect of irradiation in combination with chemotherapy or targeted therapy. Methods: The plasmid vector consisting of hypoxia response elements to regulate HSV1-TK and GFP genes, was constructed and stably transfected into human NPC cells. The expected clone was identified and validated by in vivo and in vitro assay. DSB repair was measured by γH2AX foci formation. Tumor growth delay assay and spatial biodistribution of various biomarkers was designed to investigate the anti-tumor effect. Results: The system has the propensity of high expression of reporter genes under hypoxia and low to no expression under normoxia. Intratumoral biodistributions of GFP and classic hypoxic biomarkers were identical in poor-perfused region. Upon equilibration with 10% O 2 , the xenografts showed higher expression of hypoxic biomarkers. Cisplatin radiosensitized SUNE-1/HRE cells under hypoxia by suppressing DSB repair while the addition of PI3K/mTOR inhibitor further enhanced the anti-tumoral therapeutic efficacy. Combination of IR, DDP and NVP-BEZ235 exhibited most effective anti-tumor response in vivo . These observations underline the importance of dual reporter model for imaging tumor hypoxia in therapeutic study. Conclusions: Our preclinical model enables the investigation of heterogeneous tumor hypoxic regions in xenograft tissues and explores the treatment efficacy of combinations of various therapeutic approaches to overcome hypoxia., Competing Interests: Competing Interests: The authors have declared that no competing interest exists., (© The author(s).)