Dynamic Focusing Research Articles

AMW-YOLOv8n: Road Scene Object Detection Based on an Improved YOLOv8

This study introduces an improved YOLOv8 model tailored for detecting objects in road scenes. To overcome the limitations of standard convolution operations in adapting to varying targets, we introduce Adaptive Kernel Convolution (AKconv). AKconv dynamically adjusts the convolution kernel’s shape and size, enhancing the backbone network’s feature extraction capabilities and improving feature representation across different scales. Additionally, we employ a Multi-Scale Dilated Attention (MSDA) mechanism to focus on key target features, further enhancing feature representation. To address the challenge posed by YOLOv8’s large down sampling factor, which limits the learning of small target features in deeper feature maps, we add a small target detection layer. Finally, to improve model training efficiency, we introduce a regression loss function with a Wise-IoU dynamic non-monotonic focusing mechanism. With these enhancements, our improved YOLOv8 model excels in road scene object detection tasks, achieving a 5.6 percentage point improvement in average precision over the original YOLOv8n on real road datasets.

Improved YOLOv7 Electric Work Safety Belt Hook Suspension State Recognition Algorithm Based on Decoupled Head

Safety is the eternal theme of power systems. In view of problems such as time-consuming and poor real-time performance in the correct use of seat belt hooks by manual supervision operators in the process of power operation, this paper proposes an improved YOLOv7 seat belt hook suspension state recognition algorithm. Firstly, the feature extraction part of the YOLOv7 backbone network is improved, and the M-Spatial Pyramid Pooling Concurrent Spatial Pyramid Convolution (M-SPPCSPC) feature extraction module is constructed to replace the Spatial Pyramid Pooling Concurrent Spatial Pyramid Convolution (SPPCSPC) module of the backbone network, which reduces the amount of computation and improves the detection speed of the backbone network while keeping the sensory field of the backbone network unchanged. Second, a decoupled head, which realizes the confidence and regression frames separately, is introduced to alleviate the negative impact of the conflict between the classification and regression tasks, consequently improving the network detection accuracy and accelerating the network convergence. Ultimately, a dynamic non-monotonic focusing mechanism is introduced in the output layer, and the Wise Intersection over Union (WioU) loss function is used to reduce the competitiveness of high-quality anchor frames while reducing the harmful gradient generated by low-quality anchor frames, which ultimately improves the overall performance of the detection network. The experimental results show that the mean Average Precision (mAP@0.5) value of the improved network reaches 81.2%, which is 7.4% higher than that of the original YOLOv7, therefore achieving better detection results for multiple-state recognition of hooks.

Real-Time Imaging Assessment of Stress-Induced Vascular Inflammation Using Heartbeat-Synchronized Motion Compensation.

Chronic mental stress accelerates atherosclerosis through complicated neuroimmune pathways, needing for advanced imaging techniques to delineate underlying cellular mechanisms. While histopathology, ex vivo imaging, and snapshots of in vivo images offer promising evidence, they lack the ability to capture real-time visualization of blood cell dynamics within pulsatile arteries in longitudinal studies. An electrically tunable lens was implemented in intravital optical microscopy, synchronizing the focal plane with heartbeats to follow artery movements. ApoE-/- mice underwent 2 weeks of restraint stress before baseline imaging followed by 2 weeks of stress exposure in the longitudinal imaging, while nonstressed mice remained undisturbed. The progression of vascular inflammation was assessed in the carotid arteries through intravital imaging and histological analyses. A 4-fold reduction of motion artifact, assessed by interframe SD, and an effective temporal resolution of 25.2 Hz were achieved in beating murine carotid arteries. Longitudinal intravital imaging showed chronic stress led to a 6.09-fold (P=0.017) increase in myeloid cell infiltration compared with nonstressed mice. After 3 weeks, we observed that chronic stress intensified vascular inflammation, increasing adhered myeloid cells by 2.45-fold (P=0.031), while no significant changes were noted in nonstressed mice. Microcirculation imaging revealed increased circulating, rolling, and adhered cells in stressed mice's venules. Histological analysis of the carotid arteries confirmed the in vivo findings that stress augmented plaque area, myeloid cell and macrophage accumulation, and necrotic core volume while reducing fibrous cap thickness indicating accelerated plaque formation. We visualized the 3-dimensional structure of the carotid artery and 4-dimensional dynamics of the venules in the cremaster muscle. Dynamic focusing motion compensation intravital microscopy enabled subcellular resolution in vivo imaging of blood cell dynamics in beating arteries under chronic restraint stress in real time. This novel technique emphasizes the importance of advanced in vivo imaging for understanding cardiovascular disease.

Enhancing Small Target Detection in Aerial Imagery with Bi-PAN-FPN and EDMOA-Optimized YOLOv8-s

Across the globe, people are working to build "smart cities" that will employ technology to make people's lives better and safer. Installing cameras at strategic spots across the city to monitor public spaces besides provide real-time footage to law enforcement besides other local authorities is a crucial part of smart city infrastructure, which includes video surveillance. A more effective answer is provided by deep learning algorithms, however research in this area still faces significant problems from changes in target size, form change, occlusion, and illumination circumstances as seen from the drone's perspective. In light of the aforementioned issues, this study presents a highly effective and resilient approach for aerial picture identification. To begin, the concept of Bi-PAN-FPN is presented to enhance the neck component of YOLOv8-s, taking into consideration the prevalent issue of small targets being easily misdetected or ignored in aerial photos. We achieve a more advanced and thorough feature fusion procedure much as feasible by completely considering and reusing multiscale features. To further reduce the amount of parameters in the model and prevent info loss during long-distance feature transfer, the benchmark model's backbone incorporates the GhostblockV2 structure in lieu of a portion of the C2f module. With the help of the Enhanced Dwarf Mongoose Optimization Algorithm (EDMOA), the suggested model's hyper-parameters are optimised. Lastly, a dynamic nonmonotonic focusing mechanism is employed in conjunction with WiseIoU loss as bounding box regression loss. The detector accounts for varying anchor box quality by utilizing "outlier" evaluations, thus improving the complete presentation of the detection task.

DNE-YOLO: A method for apple fruit detection in Diverse Natural Environments

The apple industry, recognized as a pivotal sector in agriculture, increasingly emphasizes the mechanization and intelligent advancement of picking technology. This study innovatively applies a mist simulation algorithm to apple image generation, constructing a dataset of apple images under mixed sunny, cloudy, drizzling and foggy weather conditions called DNE-APPLE. It introduces a lightweight and efficient target detection network called DNE-YOLO. Building upon the YOLOv8 base model, DNE-YOLO incorporates the CBAM attention mechanism and CARAFE up-sampling operator to enhance the focus on apples. Additionally, it utilizes GSConv and the dynamic non-monotonic focusing mechanism loss function WIOU to reduce model parameters and decrease reliance on dataset quality. Extensive experimental results underscore the efficacy of the DNE-YOLO model, which achieves a detection accuracy (precision) of 90.7%, a recall of 88.9%, a mean accuracy (mAP50) of 94.3%, a computational complexity (GFLOPs) of 25.4G, and a parameter count of 10.46M across various environmentally diverse datasets. Compared to YOLOv8, it exhibits superior detection accuracy and robustness in sunny, drizzly, cloudy, and misty environments, making it especially suitable for practical applications such as apple picking for agricultural robots. The code for this model is open source at https://github.com/wuhaitao2178827/DNE-YOLO.

Multi-parametric Photoacoustic Imaging Combined with Acoustic Radiation Force Impulse Imaging for Applications in Tissue Engineering.

Tissue engineering is a dynamic field focusing on the creation of advanced scaffolds for tissue and organ regeneration. These scaffolds are customized to their specific applications and are often designed to be complex, large structures to mimic tissues and organs. This study addresses the critical challenge of effectively characterizing these thick, optically opaque scaffolds that traditional imaging methods fail to fully image due to their optical limitations. We introduce a novel multi-modal imaging approach combining ultrasound, photoacoustic, and acoustic radiation force impulse imaging. This combination leverages its acoustic-based detection to overcome the limitations posed by optical imaging techniques. Ultrasound imaging is employed to monitor the scaffold structure, photoacoustic imaging is employed to monitor cell proliferation, and acoustic radiation force impulse imaging is employed to evaluate the homogeneity of scaffold stiffness. We applied this integrated imaging system to analyze melanoma cell growth within silk fibroin protein scaffolds with varying pore sizes and therefore stiffness over different cell incubation periods. Among various materials, silk fibroin was chosen for its unique combination of features including biocompatibility, tunable mechanical properties, and structural porosity which supports extensive cell proliferation. The results provide a detailed mesoscale view of the scaffolds' internal structure, including cell penetration depth and biomechanical properties. Our findings demonstrate that the developed multimodal imaging technique offers comprehensive insights into the physical and biological dynamics of tissue-engineered scaffolds. As the field of tissue engineering continues to advance, the importance of non-ionizing and non-invasive imaging systems becomes increasingly evident, and by facilitating a deeper understanding and better characterization of scaffold architectures, such imaging systems are pivotal in driving the success of future tissue-engineering solutions.

Interstitial dual-mode ultrasound with a 3-mm MR-compatible catheter for image-guided HIFU and directional in-vitro tissue ablations.

Current interstitial techniques of tumor ablation face challenges that ultrasound technologies could meet. The ablation radius and directionality of the ultrasound beam could improve the efficiency and precision. Here, a 9-gauge MR-compatible dual-mode ultrasound catheter prototype was experimentally evaluated for Ultrasound Image-guided High Intensity Focused Ultrasound (USgHIFU) conformal ablations. The prototype consisted of 64 piezocomposite linear array elements and was driven by an open research programmable dual-mode ultrasound platform. After verifying the US-image guidance capabilities of the prototype, the HIFU output performances (dynamic focusing and HIFU intensities) were quantitatively characterized, together with the associated 3D HIFU-induced thermal heating in tissue phantoms (using MR thermometry). Finally, the ability to produce robustly HIFU-induced thermal ablations in in-vitro liver was studied experimentally and compared to numerical modeling. Investigations of several HIFU dynamic focusing allowed overcoming the challenges of miniaturizing the device: mono-focal focusing maximized deep energy deposition, while multi-focal strategies eliminated grating lobes. The linear-array design of the prototype made it possible to produce interstitial ultrasound images of tissue and tumor mimics in situ. Multi-focal pressure fields were generated without grating lobes and transducer surface intensities reached up to Isapa = 14 W·cm-2. Seventeen elementary thermal ablations were performed in vitro. Rotation of the catheter proved the directionality of ablation, sparing non-targeted tissue. This experimental proof of concept demonstrates the feasibility of treating volumes comparable to those of primary solid tumors with a miniaturized USgHIFU catheter whose dimensions are close to those of tools traditionally used in interventional radiology, while offering new functionalities.

MFAD-RTDETR: A Multi-Frequency Aggregate Diffusion Feature Flow Composite Model for Printed Circuit Board Defect Detection

To address the challenges of excessive model parameters and low detection accuracy in printed circuit board (PCB) defect detection, this paper proposes a novel PCB defect detection model based on the improved RTDETR (Real-Time Detection, Embedding and Tracking) method, named MFAD-RTDETR. Specifically, the proposed model introduces the designed Detail Feature Retainer (DFR) into the original RTDETR backbone to capture and retain local details. Subsequently, based on the Mamba architecture, the Visual State Space (VSS) module is integrated to enhance global attention while reducing the original quadratic complexity to a linear level. Furthermore, by exploiting the deformable attention mechanism, which dynamically adjusts reference points, the model achieves precise localization of target defects and improves the accuracy of the transformer in complex visual tasks. Meanwhile, a receptive field synthesis mechanism is incorporated to enrich multi-scale semantic information and reduce parameter complexity. In addition, the scheme proposes a novel Multi-frequency Aggregation and Diffusion feature composite paradigm (MFAD-feature composite paradigm), which consists of the Aggregation Diffusion Fusion (ADF) module and the Refiner Feature Composition (RFC) module. It aims to strengthen features with fine-grained awareness while preserving a certain level of global attention. Finally, the Wise IoU (WIoU) dynamic nonmonotonic focusing mechanism is used to reduce competition among high-quality anchor boxes and mitigate the effects of the harmful gradients from low-quality examples, thereby concentrating on anchor boxes of average quality to promote the overall performance of the detector. Extensive experiments are conducted on the PCB defect dataset released by Peking University to validate the effectiveness of the proposed model. The experimental results show that our approach achieves the 97.0% and 51.0% performance in mean Average Precision (mAP)@0.5 and mAP@0.5:0.95, respectively, which significantly outperforms the original RTDETR. Moreover, the model reduces the number of parameters by approximately 18.2% compared to the original RTDETR.

Fast acoustic droplet ejection based on annular array transducer

Acoustic droplet ejection (ADE) has become the preferred method for liquid transfer in a variety of applications including synthetic biology, genotyping and drug discovery.Comparing with traditional pipetting techniques, the accuracy and data reproducibility of ADE based liquid transfer are improved, waste and cost are reduced, and cross-contamination is eliminated. The key component in the ADE system is the ultrasound transducer, which is responsible for generating focused ultrasound beam for droplet ejection. However, current ADE systems commonly utilize a single-element focused transducer with a fixed focal length that require mechanical movement to focus on the liquid surface, resulting in reduced liquid transfer efficiency. In this study, we first present a high-frequency annular array transducer for the ADE technology, which enables rapid and dynamic axial focusing to the liquid surface without mechanically moving the transducer, thereby accelerating liquid transfer. Experimental results show that the proposed 10 MHz, 5-element annular array transducer has good dynamic axial focusing ability, and can achieve accurate and stable droplet ejection of nanoliter volume at the designed focal length of 26–32 mm. Our results highlight the potential of the annular array transducer in advancing ADE system for rapid liquid transfer. This technology is expected to be useful in a variety of applications where precise and high-throughput liquid transfer is crucial.

Ultrasonic registration of myopic changes in the eye of Japanese quail chicks raised under the different spectrum of everyday illumination

Abstract High-resolution ultrasound techniques are employed in ophthalmology for detecting age-related changes in the eyeball’s structure, diagnosing tumours and vascular conditions. Meanwhile, acoustic microscopy methods are widely used in laboratory animal research. Our experiments utilised a short ultrasonic pulse (30-80 MHz) and B/Z-scan mode with dynamic focusing. This study utilised these parameters to measure various ocular dimensions, including the longitudinal and transverse size of the eyeball, the thickness of the lens, and the scleral-cornea, vitreous and chorioretinal complex. The sample comprised individuals of varying ages, an adult bird (65 days), and chicks across four age stages (10, 23, 36, 45 days), all studied in situ. The method enabled observation of age-related alterations in ocular structures throughout the ontogenesis of the animal, illuminated daily using an incandescent lamp alongside yellow and blue light filters. The outcomes illustrate that the measurement technique’s sensitivity detects an anterior-posterior axis (APA) escalation of quail eyes with an error range of ±15 microns. The method’s sensitivity adequately analysed the impact of the spectral composition of everyday lighting on deviations in eye development.

Line-field confocal optical coherence tomography based on tandem interferometry with a focus-tunable lens.

This article introduces an innovative line-field confocal optical coherence tomography (LC-OCT) system based on tandem interferometry, featuring a focus-tunable lens for dynamic focusing. The principle of tandem interferometry is first recalled, and an analytical expression of the interferometric signal detected is established in order to identify the influence of key experimental parameters. The LC-OCT system is based on a Linnik-type imaging interferometer with a focus-tunable lens for focus scanning, coupled to a Michelson-type compensating interferometer using a piezoelectric linear translation stage for coherence plane scanning. The system achieves axial and lateral image resolutions of approximately 1 µm over the entire imaging depth (400 µm), in line with conventional LC-OCT. Vertical section images (B-scans) of skin acquired at 14.3 fps reveal distinguishable structures within the epidermis and dermis. Using refocusing and stitching, images of a tissue phantom were obtained with an imaging depth superior to 1.4 mm. The system holds promise for LC-OCT miniaturization, along with enhanced imaging speed and extended imaging depth.

Ultrasonic evaluation of wire arc additively manufactured parts with heterogeneous microstructure

Abstract In Wire Arc Additive Manufacturing (WAAM) components manufactured by gas metal arc welding strategy, periodically alternating patterns of fine- and coarse-grained microstructures develop along the building direction, depending on the temperature evolution. Ultrasonic non-destructive evaluation relies on the scattering of waves from defects. In the material microstructure, ultrasound waves are also scattered at the grain boundaries as grain noise. The work presents a study on the influence of heterogeneous microstructure in WAAM on defect detectability during ultrasonic inspections. The study used a 2D finite element model to simulate ultrasound wave propagation in the material microstructure represented by a surface element from electron backscatter diffraction scans. Side-drilled holes were introduced in the material as defects. Phased array was used to perform full matrix capture and obtain ultrasound image through total focusing method, with dynamic focusing capability to inspect the heterogeneous microstructure. The study analysed the defect echoes and noises using the signal-to-noise ratio (SNR) to assess its defect detectability. While defects were well detected, the SNR values differ by 10% for defects at various positions along the building and transverse directions in the heterogeneous microstructure.

Dynamic near-field and far-field radiation manipulation using a reprogrammable guided-wave-excited metasurface

The dynamic and integrated control of near- and far-field electromagnetic waves is essential for advancing emerging intelligent information technology. Metasurfaces, distinguished by their low-profile design, cost-effectiveness, and ease of fabrication, have successfully revolutionized various electromagnetic functions. However, current research on the dynamic integrated manipulation of near-field and far-field electromagnetic waves using a single metasurface remains relatively constrained, due to the complexity of element-level control, restricted dynamic tuning range, and tuning speed. Herein, we propose an element-level controlled, versatile, compact, and broadband platform allowing for the real-time electronic reconstruction of desired near/far-field electromagnetic wavefronts. This concept is achieved by precisely regulating the 1-bit amplitude coding pattern across a guided-wave-excited metasurface aperture loaded with PIN diodes, following our binary-amplitude holographic theory and modified Gerchberg–Saxton (G–S) algorithm. Consistent findings across calculations, simulations, and experiments highlight the metasurface’s robust performance in 2D beam scanning, frequency scanning, dynamic focusing lens, dynamic holography display, and 3D multiplexing holography, even under 1-bit control. This simplified and innovative metasurface architecture holds the promise of substantially propelling forthcoming investigations and applications of highly integrated, multifunctional, and intelligent platforms.

Optical design and analysis of a high-speed triple galvanometer laser 3D scanning system

In this paper, a triple galvanometer laser three-dimensions (3D) scanning system (TGLSS) was proposed. The system consisted of three galvanometers and two parabolic mirrors, with rapid movement of the laser focus in 3D space by rotations of the three galvanometers. The optical principle of the TGLSS was analyzed, and the method of designing TGLSS by geometrical optics and Abbe’s sine condition was given in detail. Scanning range and focusing quality were analyzed by optical simulations. An experiment was carried out to verify the feasibility of TGLSS. The implications of TGLSS for reflective dynamic focusing device (RDFD)-based laser system and their impact on technologically relevant applications were discussed. The great potential for enhancing the speed of advanced laser processing in 3D space was offered by TGLSS.

Characterization and functional analysis of chicken promyelocytic leukemia protein

In mammals, promyelocytic leukemia (PML) protein, also named as TRIM19, is the key component of nuclear membrane-less sub structures PML nuclear bodies (PML-NBs) or nuclear domains 10 (ND10). PML-NBs are dynamic foci that consist of numerous permanently or transiently associated proteins. The mammalian PMLs are involved in the regulation of various cellular pathways, including apoptosis, intrinsic and innate antiviral immunity, cell cycle, DNA damage, senescence and etc. Nevertheless, little is known about the role of chicken PML (chPML). In this study, chPML gene was cloned, and its several functions were characterized. We found that chPML was widely expressed in different tissues of chickens, and showed different subcellular distribution pattern in DF-1 cells comparing with LMH and HD11 cells. Like human PML, chPML was identified to be SUMOylated. K463 is one critical SUMOylation site and 240RARRG244 is SUMO interaction motif (SIM) of chPML. Moreover, qPCR showed that chPML could not only up-regulate the expression of host innate immune factor IFN-β and its downstream ISGs, but also antigen presentation-related factors including class II transactivator (CIITA) and MHC II DM beta 2 (DMB2). Notably, over-expression of chIFN-β could promote the expression of endogenous chPML. All these provide novel insights into the function of chPML, and pave the way for further studying the roles of chPML in biological process and anti-infection function.

ESFD-YOLOv8n: Early Smoke and Fire Detection Method Based on an Improved YOLOv8n Model

Ensuring fire safety is essential to protect life and property, but modern infrastructure and complex settings require advanced fire detection methods. Traditional object detection systems, often reliant on manual feature extraction, may fall short, and while deep learning approaches are powerful, they can be computationally intensive, especially for real-time applications. This paper proposes a novel smoke and fire detection method based on the YOLOv8n model with several key architectural modifications. The standard Complete-IoU (CIoU) box loss function is replaced with the more robust Wise-IoU version 3 (WIoUv3), enhancing predictions through its attention mechanism and dynamic focusing. The model is streamlined by replacing the C2f module with a residual block, enabling targeted feature extraction, accelerating training and inference, and reducing overfitting. Integrating generalized efficient layer aggregation network (GELAN) blocks with C2f modules in the neck of the YOLOv8n model further enhances smoke and fire detection, optimizing gradient paths for efficient learning and high performance. Transfer learning is also applied to enhance robustness. Experiments confirmed the excellent performance of ESFD-YOLOv8n, outperforming the original YOLOv8n by 2%, 2.3%, and 2.7%, with a mean average precision (mAP@0.5) of 79.4%, precision of 80.1%, and recall of 72.7%. Despite its increased complexity, the model outperforms several state-of-the-art algorithms and meets the requirements for real-time fire and smoke detection.

Multi-layered medium ultrasonic phased array sparse TFM imaging based on self-adaptive differential evolution algorithm

Abstract Multilayer Composite material structures have been widely used in modern engineering fields. However, defects within these materials can adversely affect mechanical properties. Ultrasonic phased array total focusing method (TFM) imaging has advantages of high precision and dynamic focusing over the entire range, achieving significant progress in homogeneous medium detection. However, heavy computational burdens of multilayer structures lead to inefficient imaging. To address this issue, a sparse-TFM imaging algorithm using ultrasonic phased arrays suitable for multilayer media is proposed in this paper. This method constructs a fitness function with constraints such as main lobe width and sidelobe peak. Its objective is to obtain the distribution of sparse array element positions using an self-adaptive differential evolution algorithm. Subsequently, the delay time of each array element in multilayer media sparse TFM is calculated using the root mean square (RMS) principle and combined with amplitude weighting, the method corrects the imaging results. Compared with the Ray-based full-matrix capture and TFM method (Ray-based FMC/TFM), the RMS-based full-matrix capture and TFM (RMS-based FMC/TFM), and the phase shift method, the experimental and simulation results demonstrate that the proposed method significantly reduces the imaging data volume, improves computational efficiency, and maintains quantitative errors within 0.2 mm.

An improved YOLOv5-based apple leaf disease detection method

The effective identification of fruit tree leaf disease is of great practical significance to reduce pesticide spraying, improve fruit yield and realize ecological agriculture. Computer vision technology can be effectively identifying and prevent plant diseases and insect pests. However, the lack of consideration of disease diversity and accuracy of existing detection models hinders their application and development in the field of plant pest detection. This paper proposes an efficient detection model of apple leaf disease spot through the improvement of the traditional Yolov5 detection network called A-Net. In order to significantly increase the A-Net's detection speed and accuracy, the A-Net model applies the loss function Wise-IoU, which includes the attention mechanism and the dynamic focusing mechanism, to the Yolov5 network model. The RepVGG module is then used to replace the original model's convolution module. The experimental results show that the improved model effectively suppresses the growth of some error weights. Compared with several object detection models, the improved A-Net model has a Mean Average Precision across IoU threshold 0.5 and an accuracy of 92.7%, which fully proves that the improved A-Net model has more advantages in detecting apple leaf diseases.

Automatic Perception of Typical Abnormal Situations in Cage-Reared Ducks Using Computer Vision.

Overturning and death are common abnormalities in cage-reared ducks. To achieve timely and accurate detection, this study focused on 10-day-old cage-reared ducks, which are prone to these conditions, and established prior data on such situations. Using the original YOLOv8 as the base network, multiple GAM attention mechanisms were embedded into the feature fusion part (neck) to enhance the network's focus on the abnormal regions in images of cage-reared ducks. Additionally, the Wise-IoU loss function replaced the CIoU loss function by employing a dynamic non-monotonic focusing mechanism to balance the data samples and mitigate excessive penalties from geometric parameters in the model. The image brightness was adjusted by factors of 0.85 and 1.25, and mainstream object-detection algorithms were adopted to test and compare the generalization and performance of the proposed method. Based on six key points around the head, beak, chest, tail, left foot, and right foot of cage-reared ducks, the body structure of the abnormal ducks was refined. Accurate estimation of the overturning and dead postures was achieved using the HRNet-48. The results demonstrated that the proposed method accurately recognized these states, achieving a mean Average Precision (mAP) value of 0.924, which was 1.65% higher than that of the original YOLOv8. The method effectively addressed the recognition interference caused by lighting differences, and exhibited an excellent generalization ability and comprehensive detection performance. Furthermore, the proposed abnormal cage-reared duck pose-estimation model achieved an Object Key point Similarity (OKS) value of 0.921, with a single-frame processing time of 0.528 s, accurately detecting multiple key points of the abnormal cage-reared duck bodies and generating correct posture expressions.

Liquid Crystal Integrated Dynamic Terahertz Metalenses

AbstractMetasurfaces can freely manipulate the wavefront of electromagnetic waves with the merits of minimization and integration. Dynamic metadevices have continued to be urgently pursued to realize on‐demand functionalities. Among them, dynamic metalenses have attracted particular attention due to their widespread applications, especially in the promising terahertz region. Here, a liquid crystal integrated terahertz metalens with dynamic focusing properties is proposed. Subwavelength silicon columns are specifically designed and set to realize the same lens phase profile for orthogonal linear polarizations, while the rotation of columns additionally introduces a spin‐selective conjugated geometric phase. A photopatterned liquid crystal layer is added to compensate for the geometric phase; therefore, a polarization‐independent single focus is achieved. With a saturated bias applied to the liquid crystal, the phase compensation vanishes; therefore, a spin‐selective bifocal function is realized. When lens and polarization grating profiles are encoded to the geometric phase separately, longitudinal and lateral spin‐dependent focus separations are demonstrated. The metalens can be electrically switched between single‐focus and bifocal functions. The dynamic metalens can be utilized in depth‐of‐field imaging and telecom port selection. The design endows various existing metadevices with dynamic functions.