Application and dosimetric comparison of surface-guided deep inspiration breath-hold for lung stereotactic body radiotherapy.

Respiratory motion management is the crucial challenge for safe and effective application of lung stereotactic body radiotherapy (SBRT). The present study implemented lung SBRT treatment in voluntary deep inspiration breath-hold (DIBH) with surface-guided radiotherapy (SGRT) system and evaluated the geometric and dosimetric benefits of DIBH to organs-at-risk (OARs), aiming to advising the choice between DIBH technology and conventional free breathing 4 dimensions (FB-4D) technology. Five patients of lung SBRT treated in DIBH with SGRT at our institution were retrospectively analyzed. CT scans were acquired in DIBH and FB-4D, treatment plans were generated for both respiratory phases. The geometric and dosimetry of tumor, ipsilateral lung, double lungs and heart were compared between the DIBH and FB-4D treatment plans. In terms of target coverage, utilizing DIBH significantly reduced the mean plan target volume (PTV) by 21.9% (p = 0.09) compared to FB-4D, the conformity index (CI) of DIBH and FB-4D were comparable, but the dose gradient index (DGI) of DIBH was higher. With DIBH expanding lung, the volumes of ipsilateral lung and double lungs were 2535.1 ± 403.0cm3 and 4864.3 ± 900.2cm3, separately, 62.2% (p = 0.009) and 73.1% (p = 0.009) more than volumes of ipsilateral lung (1460.03 ± 146.60cm3) and double lungs (2811.25 ± 603.64cm3) in FB-4D. The heart volume in DIBH was 700.0 ± 146.1cm3, 11.6% (p = 0.021) less than that in FB-4D. As for OARs protection, the mean dose, percent of volume receiving > 20Gy (V20) and percent of volume receiving > 5Gy (V5) of ipsilateral lung in DIBH were significantly lower by 33.2% (p = 0.020), 44.0% (p = 0.022) and 24.5% (p = 0.037) on average, separately. Double lungs also showed significant decrease by 31.1% (p = 0.019), 45.5% (p = 0.024) and 20.9% (p = 0.048) on average for mean dose, V20 and V5 in DIBH. Different from the lung, the mean dose and V5 of heart showed no consistency between DIBH and FB-4D, but lower maximum dose of heart was achieved in DIBH for all patients in this study. Appling lung SBRT in DIBH with SGRT was feasibly performed with high patient compliance. DIBH brought significant dosimetric benefits to lung, however, it caused more or less irradiated heart dose that depend on the patients' individual differences which were unpredictable.

A Glove-Wearing Detection Algorithm Based on Improved YOLOv8.

Wearing gloves during machinery operation in workshops is essential for preventing accidental injuries, such as mechanical damage and burns. Ensuring that workers are wearing gloves is a key strategy for accident prevention. Consequently, this study proposes a glove detection algorithm called YOLOv8-AFPN-M-C2f based on YOLOv8, offering swifter detection speeds, lower computational demands, and enhanced accuracy for workshop scenarios. This research innovates by substituting the head of YOLOv8 with the AFPN-M-C2f network, amplifying the pathways for feature vector propagation, and mitigating semantic discrepancies between non-adjacent feature layers. Additionally, the introduction of a superficial feature layer enriches surface feature information, augmenting the model's sensitivity to smaller objects. To assess the performance of the YOLOv8-AFPN-M-C2f model, this study conducted multiple experiments using a factory glove detection dataset compiled for this study. The results indicate that the enhanced YOLOv8 model surpasses other network models. Compared to the baseline YOLOv8 model, the refined version shows a 2.6% increase in mAP@50%, a 63.8% rise in FPS, and a 13% reduction in the number of parameters. This research contributes an effective solution for the detection of glove adherence.

New Progress in Artificial Intelligence Algorithm Research Based on Big Data Processing of IOT Systems on Intelligent Production Lines.

Intelligent production line is the abbreviation of intelligent production line. Intelligent production line refers to a form of production organization that uses intelligent manufacturing technology to realize the production process of products. The actual manufacturing process includes different levels and links, and each step cooperates to create a high-efficiency production line. The intelligent production line includes 3 levels covering automation equipment, digital workshops, and intelligent factories and runs through 6 major links of intelligent manufacturing (intelligent management, intelligent monitoring, intelligent processing, intelligent assembly, intelligent inspection, and intelligent logistics). The emergence of the Internet of Things system has changed the way of information dissemination. The system combines radio frequency automatic identification and global positioning system technologies to achieve functions such as information exchange and processing, enabling information processing to be intelligent and improving resource utilization. Big data processing includes multiple data processing procedures, but data quality is the most important link in the entire process, and each data processing link will have an impact on the quality of big data. The big data processing process mainly includes data collection, data preprocessing, data storage, data processing and analysis, data display, data visualization, data application, and other links. This article aims to study the new progress of artificial intelligence algorithms for big data processing of IOT systems on intelligent production lines. It is hoped that through the development of intelligent production lines and big data processing technologies, ways to optimize artificial intelligence algorithms can be found. This study proposes a metadata replication method based on a separate replication strategy, which separates the replication process of the data operation log, each is independent, and shortens the data replication time. Combining the existing intelligent production line network platform in the laboratory and carrying out the research of the intelligent production line network state prediction system based on the neural network to design a network prediction system can prejudge the operation status of the intelligent production line network. The experimental results in this article show that when the Namenode mode is used to read data and when the number of clients reaches 8, the data processing basically remains unchanged. When the NCluster system reads data and when the number of clients is 6, the data is processed 1256. When the number of clients is 20, the data is processed 2100, the NCluster system will remain stable when the number of clients reaches 12, and compared with the Namenode system, it has obvious advantages.

Numerical study of natural convection in a horizontal cylinder filled with water-based alumina nanofluid.

Natural heat convection of water-based alumina (Al2O3/water) nanofluids (with volume fraction 1% and 4%) in a horizontal cylinder is numerically investigated. The whole three-dimensional computational fluid dynamics (CFD) procedure is performed in a completely open-source way. Blender, enGrid, OpenFOAM and ParaView are employed for geometry creation, mesh generation, case simulation and post process, respectively. Original solver 'buoyantBoussinesqSimpleFoam' is selected for the present study, and a temperature-dependent solver 'buoyantBoussinesqSimpleTDFoam' is developed to ensure the simulation is more realistic. The two solvers are used for same cases and compared to corresponding experimental results. The flow regime in these cases is laminar (Reynolds number is 150) and the Rayleigh number range is 0.7 × 10(7) ~ 5 × 10(7). By comparison, the average natural Nusselt numbers of water and Al2O3/water nanofluids are found to increase with the Rayleigh number. At the same Rayleigh number, the Nusselt number is found to decrease with nanofluid volume fraction. The temperature-dependent solver is found better for water and 1% Al2O3/water nanofluid cases, while the original solver is better for 4% Al2O3/water nanofluid cases. Furthermore, due to strong three-dimensional flow features in the horizontal cylinder, three-dimensional CFD simulation is recommended instead of two-dimensional simplifications.