Occlusion Conditions Research Articles

Construction of tennis pose estimation and action recognition model based on improved ST-GCN

With the rapid growth of computer vision and deep learning technologies, the application of pose estimation and action recognition in sports training has become increasingly widespread. Due to factors such as complex movements, fast speed, and limb occlusion, pose estimation and action recognition in tennis face significant challenges. Therefore, this study first introduces selective dropout and pyramid region of interest pooling layer strategies in fast region convolutional neural networks. Secondly, a pose estimation algorithm based on multi-scale fusion pose residual network 50 is designed, and finally a spatiotemporal graph convolutional network model is constructed by fusing channel attention module and multi-scale dilated convolution module. The data showed that the average detection accuracy of the improved attitude residual network 50 was 70.4%, and the accuracy of object detection for small, medium, and large objects was 57.4%, 69.3%, and 79.2%, respectively. The continuous action recognition accuracy and inter action fluency detection time of the improved spatiotemporal graph convolutional network were 93.8% and 19.2 ms, respectively. When the sample size was 1000, its memory usage was 1378 MB and the running time was 32.7 ms. Experiments have shown that the improved model achieves high accuracy and robustness in tennis action recognition tasks, especially in complex scenes and limb occlusion conditions, where the model shows significant advantages. This study aims to provide an efficient and accurate motion recognition technology for tennis posture analysis and intelligent training.

Survey on the management of retinal arterial occlusions in Spain

ObjectiveTo determine the current status of retinal artery occlusion (RAO) management in Spain. Material and MethodsAn anonymous questionnaire of 22 questions was carried out on the Google Forms® platform between April 27 and May 28, 2023. This questionnaire was distributed via email by the Spanish Society of Retina and Vitreous and the Spanish Society of Ophthalmology. Results159 ophthalmologists from 91 public and 44 private centers in the 17 autonomous communities answered the questionnaire. Fifty-one percent of the ophthalmologists who answered the survey were women, 56.6% worked in public centers, 9.4% in private centers and 34% in both centers.Of the ophthalmologists working in public centers, 41.7% attend ophthalmologic emergencies 24 hours a day, 365 days a year, while the rest delay their attention outside the usual hours. In 62.6% of the public centers with responses collected there are no protocols for the management of OAR in the ophthalmology services and as for multidisciplinary protocols, 57% recognize that they do not have one, 23% have one with neurology or internal medicine and only 13% of the centers include OAR within the stroke code. The most frequently reported treatments are hypotensors and anterior chamber paracentesis.In private centers, 73.7% report not having their own protocol, and 78.9% do not have a protocol with other services.The main barriers perceived by the ophthalmologists surveyed were: patients arriving outside the window period (77.2%), no specific treatment and lack of established protocols (55.4%), low incidence (32.9%).Suggestions for improvement were: need to implement protocols, coordination with other services and that OAR be included as a stroke code. ConclusionsDespite its limitations, (simple questionnaire, open design of some questions, lack of representation of centers and provinces and contradictory responses within the same center), we consider that it can serve as an approximation to the current management of RAO in Spain. It is necessary to develop protocols adapted to the different centers and to comply with them effectively in order to improve accessibility and equity.

Removable denture use, fit, and all-cause mortality in older adults with reduced occlusal support: The OHSAKA study

Statement of problemAs the global population ages, tooth loss among older adults has become increasingly prevalent, impacting oral function and overall health. Although removable dentures have been used to improve masticatory function, the long-term effects of denture conditions on mortality are unclear. PurposeThe purpose of this clinical study was to clarify the association between removable denture condition and the mortality rate in older adults with reduced occlusal support. Material and methodsThis longitudinal study analyzed survival data (up to 2022) for 186893 individuals aged ≥75 years who underwent dental examinations through public healthcare services between 2018 and 2020. The participants were categorized into 10 groups (A1 to A3, B1 to B4, and C1 to C3) based on their occlusal support status using the Eichner index. Denture conditions were assessed by private practice dentists and classified into 3 categories: well-fitting, poorly fitting, and not used. Cox proportional hazards regression analyses were conducted to examine the association between denture conditions and mortality by considering the following control variables: age, sex, history of chronic diseases (hypertension, diabetes, hyperlipidemia, and dementia), smoking habits, and body mass index for the entire study population and within each of the B1 to B4 and C1 to C3 groups. ResultsSignificantly higher hazard ratios (HRs) for mortality were observed with occlusal support deterioration, poorly fitting dentures, and nonusage, with the peak observed in the C3 group with no dentures. The HRs (with 95% confidence intervals) of well-fitting dentures were consistently the lowest across all groups as follows: 0.98 (0.90 to 1.06) in B1, 1.02 (0.95 to 1.10) in B2, 1.11 (1.03 to 1.19) in B3, 1.14 (1.07 to 1.22) in B4, 1.15 (1.05 to 1.26) in C1, 1.28 (1.21 to 1.37) in C2, and 1.58 (1.46 to 1.72) in C3, with the A1 to A3 group as the reference. Poorly fitting or no dentures were associated with increased HRs in the Eichner groups B3 to C3, compared with well-fitting dentures. ConclusionsThis large population-based cohort study found that well-fitting dentures were associated with lower mortality risk in older adults with reduced posterior occlusal support. Nonusage or poorly fitting dentures were associated with an increased risk of mortality as occlusal support deteriorated.

A multimodal framework for pepper diseases and pests detection

Pepper diseases and pests typically exhibit small target proportions, diverse shapes and sizes, complex imaging backgrounds, and similarities with the background. Existing detection methods perform poorly in identifying targets of different sizes and shapes within the same scene, and they lack adequate noise suppression capabilities. To address the practical needs of detecting pepper diseases and pests in complex scenarios, we have constructed the first multimodal pepper diseases and pests object detection dataset (PDD). This dataset includes a wide variety of diseases and pests images, along with detailed natural language descriptions of their attributes. Locating the described targets in complex scenes with similar disease symptoms and leaf occlusion presents a significant challenge. To tackle this issue, we propose the PepperNet model for object detection in pepper diseases and pests images using natural language descriptions. This model decomposes complex multimodal features of language and images into explicit attribute features and employs fine-grained multimodal attribute contrast learning strategies. This approach effectively distinguishes subtle local differences between similar objects, achieving fine-grained mapping from language to vision in complex scenarios. Our detection results show a mAP@0.5 of 91.93% and a detection speed of 121.8 frames per second. Visualizations indicate that the model maintains high robustness under varying noise levels and occlusion conditions, demonstrating superior performance and stability across diverse complex scenarios.

Safety and Efficacy of Tiny (≤3 mm) Unruptured Middle Cerebral Artery Aneurysm Treatment: An Analysis of the NeuroVascular Quality Initiative-Quality Outcomes Database Cerebral Aneurysm Registry.

Improved imaging modalities have led to an increased detection of intracranial aneurysms, many of which are small. There is uncertainty in the appropriate management of tiny aneurysms. The objective of this study was to use a large, multi-institutional NeuroVascular Quality Initiative-Quality Outcomes Database (NVQI-QOD) to assess the frequency, safety, and efficacy of treatment of tiny, unruptured middle cerebral artery (MCA) aneurysms. The NeuroVascular Quality Initiative-Quality Outcomes Database registry was queried for patients with tiny unruptured MCA aneurysms who underwent treatment. Tiny size was defined as an aneurysm with a maximum dimension of ≤3 mm. Demographics, aneurysm characteristics, and treatment safety were queried. Outcomes included modified Rankin Score (mRS) at discharge and the last follow-up as well as aneurysm occlusion status at discharge. Of 674 treated, unruptured MCA aneurysms, 57 (8.5%) were tiny. The mean aneurysm width was 2.2 mm, and the mean patient age was 55.9 years. Most aneurysms were treated with microsurgery (61.4%, 35/57). The overall intraoperative complication rate was 5.3% (3/57), and the postoperative complication rate was 10.5% (6/57). 10.5% (6/57) of patients were discharged to rehabilitation. At discharge, 42 (87.5%) of the treated aneurysms had complete occlusion. In the subgroup of patients with recorded follow-up data, 48.3% (14/29) had a mRS of 0 at discharge and 46.9% (15/32) had an mRS of 0 at the last follow-up (median follow-up 166 days). Treatment of tiny, unruptured MCA aneurysms is efficacious but may have a high rate of complications. Physicians should be cautious when deciding to treat tiny, unruptured MCA aneurysms.

Longitudinal association between posterior occlusal contact and dementia development in a large retrospective cohort study using a Japanese claims database

This study examines the association between posterior occlusal contact and the risk of dementia development in the Japanese population, utilizing Eichner classification to evaluate occlusal status. Data from Japanese health insurance claims were analyzed for the period from April 2016 to March 2022. Participants had undergone specific health checkups, had no prior history of dementia, and were classified according to their dental occlusal contact. Dementia diagnoses were determined using ICD-10 codes, and participants were divided into three groups—A, B, and C—based on the Eichner classification, which indicates the extent of occlusal contact. Over an average follow-up period of 35.6 months, 691 dementia were identified among 931,309 participants. Those diagnosed with dementia were more likely to belong to Eichner B and C groups, signifying reduced occlusal contact. After adjusting for covariates, the hazard ratios (95% confidence intervals) for Eichner B and C were 1.73 (1.31–2.28) and 2.10 (1.35–3.26), respectively. Sensitivity analyses confirmed these findings in adults aged 60–75. These findings suggest that reduced posterior occlusal contact correlates with an increased risk of dementia. Since the study is limited to participants under the age of 75, further research is required to determine its generalizability to older populations.

Method for Non-Contact Measuring the Weight of Sturgeon in Intensive Aquaculture

Weight information plays a pivotal role in sturgeon breeding and production management. However, manual measurement is time consuming and labor intensive due to the immense size of the sturgeon. Due to the unique body shape of the sturgeon, traditional image segmentation algorithms struggle to extract the necessary features from sturgeon images, which makes them unsuitable for this particular species. Moreover, accurately measuring weight in an occlusion environment is difficult. To address these challenges, an improved YOLOv5s model with a context augmentation module, focal-efficient intersection over union, and soft non-maximum suppression was proposed in this paper. To validate the model’s feasibility, the improved YOLOv5s model was first pre-trained using the sturgeon dataset, followed by further training on the occlusion dataset for segmentation tasks. Based on the phenotypic data obtained from the improved model, a multilayer perceptron method was used to estimate the sturgeon’s weight accurately. Experimental results demonstrated that the average precision of the improved YOLOv5s model reached 89.80% under occlusion conditions, and the correlation coefficient of noncontact weight measurement results reached 89.80%. The experimental results showed that the improved algorithm effectively performs segmentation of sturgeon in occlusion conditions and can accurately estimate the mass.

Oral function status of older patients seeking dental implant treatment.

In recent years, dental implant treatment has become an option for prosthetic treatment for missing teeth and is often performed in older patients. However, the complex oral functional decline in old age presents challenges in terms of frailty prevention, making oral function management after prosthetic treatment crucial. Nonetheless, the actual status of oral function in older patients seeking dental implant treatment remains unclear. In this study, we aimed to assess the oral function status of older patients seeking dental implant treatment. Among patients receiving prosthetic treatment for missing teeth, 227 older patients (111 in the pre-dental implant group and 116 in the pre-bridge/denture group) who underwent a thorough examination of their oral function were included in this study. Age, sex, comorbidities, smoking status, number of functional teeth, and occlusal support status were obtained from the medical records. Statistical analyses were performed using the t-test, chi-square test, and logistic regression (p < 0.05). Compared with the pre-bridge/denture group, the pre-dental implant group had significantly better oral hygiene, occlusal force, tongue-lip motor function, tongue pressure, masticatory function, and swallowing function and a significantly lower prevalence of oral hypofunction. Older age and decreased occlusal support were associated with the diagnosis of oral hypofunction, even after adjusting for confounding factors including prosthetic treatment. Although older patients seeking dental implant treatment have a higher oral function than those seeking general prosthetic treatment, older age and a lower number of occlusal supports suggest that appropriate oral function management is needed.

Open reduction and internal fixation of mandibular condyle fractures using AUTOFIX 2.0® cannulated compression screw system: A technical innovation with intraoperative CT-scan

Condylar process fractures remain a matter of ongoing controversy in maxillofacial surgery because of variety of opinions and proposed treatment modalities offered in the literature. The trend is toward open reduction and internal fixation (ORIF) whenever there is displacement or dislocation combined with unstable occlusal conditions. The fundamental treatment goals are anatomical fixation and early return to function. However, ORIF of a fractured condyle carries significant risks, including facial nerve damage, and presents challenges in reducing and stabilizing the fractured segment in an area of limited access. We present here an innovative technique for ORIF of condylar head fractures. It consists in fracture reduction and stabilization with two Kirschner wires followed by osteosynthesis with two cannulated headless compression screws using the AUTOFIX 2.0® system (Stryker, Kalamazoo, USA). This new technique, combined with intraoperative CT scan, ensures satisfactory reduction and stabilization of the dislocated joint segment, thereby restoring the condyle to its physiologically normal position.

Recognizing objects in complex scenes: A Recent Systematic Review

This systematic review provides a comprehensive examination of recent advancements in object recognition within complex scenes, focusing on addressing challenges such as clutter, occlusion, and diverse environmental conditions. Leveraging the PRISMA framework, the study meticulously analysed a diverse range of literature from esteemed sources, employing advanced search methods on Scopus and WoS databases to identify and analyse primary research studies (n = 25). The review encompasses three key themes: Theme 1 concentrates on Image Noise Identification and Removal Techniques, while Theme 2 delves into Image Classification and Recognition under Noise. Additionally, Theme 3 explores Innovative Models and Approaches for Noise-Robust Image Analysis. Despite the progress achieved, contemporary recognition systems struggle with real-world complexities such as varied scales, lighting variations, and different viewpoints. The synthesis of findings emphasizes the necessity for innovative strategies that capitalize on contextual cues and harness the potential of deep learning to enhance precision in object recognition within intricate visual environments. The insights gleaned from this synthesis are poised to guide future research directions, informing the development of more resilient algorithms capable of navigating challenges and catalysing advancements in the field of object recognition.

Tooth eruption status and bite force determine dental microwear texture gradients in albino rats (Rattus norvegicus forma domestica).

Dental microwear texture analysis (DMTA) is widely applied for inferring diet in vertebrates. Besides diet and ingesta properties, factors like wear stage and bite force may affect microwear formation, potentially leading to tooth position-specific microwear patterns. We investigated DMTA consistency along the upper cheek tooth row in young adult female rats at different growth stages, but with erupted adult dentitions. Bite forces for each molar (M) position were determined using muscle cross-sectional areas and lever arm mechanics. Rats were categorized into three size classes based on increasing skull length. Maximum bite force increased with size, while across all size classes, M3 bite force was almost 1.4 times higher than M1 bite force. In size class 1, M1 and M2 showed higher values than M3 for DMTA complexity, height, and volume parameters, while in size class 3, M1 had the lowest values. Comparing the same tooth position between size classes revealed opposing trends: M1 and M2 showed, for most parameters, decreasing roughness and complexity from size class 1-3, while M3 displayed the opposite trend, with size class 1 showing lowest, and either size class 2 or 3 the highest roughness and complexity values. This suggests that as rats age and M3 fully occludes, it becomes more utilized during mastication. DMTA, being a short-term diet proxy, is influenced by eruption and occlusion status changes. Our findings emphasize the importance of bite force and ontogenetic stage when interpreting microwear patterns and advise to select teeth in full occlusion for diet reconstruction.

Woven EndoBridge device apposition and compression using Sim&Size virtual simulation correlate with aneurysm occlusion status: a retrospective cohort study

BackgroundVirtual simulation is increasingly used for aneurysm treatment. This study aimed to explore whether mechanical behavior biomarkers of the Woven EndoBridge (WEB) device as computed by Sim&Size simulation software were...

A Two-Stream Method for Human Action Recognition Using Facial Action Cues.

Human action recognition (HAR) is a critical area in computer vision with wide-ranging applications, including video surveillance, healthcare monitoring, and abnormal behavior detection. Current HAR methods predominantly rely on full-body data, which can limit their effectiveness in real-world scenarios where occlusion is common. In such situations, the face often remains visible, providing valuable cues for action recognition. This paper introduces Face in Action (FIA), a novel two-stream method that leverages facial action cues for robust action recognition under conditions of significant occlusion. FIA consists of an RGB stream and a landmark stream. The RGB stream processes facial image sequences using a fine-spatio-multitemporal (FSM) 3D convolution module, which employs smaller spatial receptive fields to capture detailed local facial movements and larger temporal receptive fields to model broader temporal dynamics. The landmark stream processes facial landmark sequences using a normalized temporal attention (NTA) module within an NTA-GCN block, enhancing the detection of key facial frames and improving overall recognition accuracy. We validate the effectiveness of FIA using the NTU RGB+D and NTU RGB+D 120 datasets, focusing on action categories related to medical conditions. Our experiments demonstrate that FIA significantly outperforms existing methods in scenarios with extensive occlusion, highlighting its potential for practical applications in surveillance and healthcare settings.

Safe Endovascular Recanalization of Internal Carotid Artery Occlusion Using Retrograde Aspiration Angiography

ObjectsCerebral infarction, resulting from acute internal carotid artery (ICA) occlusion, typically manifests with a wide ischemic area and severe symptoms. Mechanical thrombectomy proves effective within 24 hours of disease onset and in less ischemic infarction core. However, in cases with well-developed collateral flow and mild symptoms, conservative treatment is initially selected. If symptoms worsen or ipsilateral hemisphere perfusion decreases, superficial temporal artery (STA) – middle cerebral artery (MCA) anastomosis is considered. Revascularization therapy may also be effective. Contralateral angiography allows for ipsilateral blood flow verification once complete revascularization is achieved, albeit with potential treatment-related complications. Here, we describe retrograde angiography using an intermediate catheter (IMC) and contralateral contrast injection to achieve safer revascularization. MethodsWe present ten acute ICA occlusion cases caused by thromboembolism, arteriosclerosis, carotid artery stent occlusion, or embolic stroke of undetermined source. Employing bilateral femoral artery puncture, one catheter was guided to the occluded ICA, while another was positioned for diagnostic purposes in the contralateral ICA or vertebral artery. The IMC was navigated through the occluded vessel, with contralateral imaging performed during aspiration. The catheter was gradually retracted to confirm the extent of occlusion and recognize recanalization. ResultsRecanalization was successfully achieved in all cases. Symptomatic ischemic complication and hyperperfuison was not seen in all cases. No cases were decreased modified Rankin scale after treatment. ConclusionAdvancements in IMC technology facilitate distal catheter guidance to the ICA, even in proximal occlusion. Combining this technique with contralateral imaging allows us to confirm the extent of occlusion and recanalization status without antegrade imaging, making recanalization therapy safer.

Nonlinear Finite Element Analysis of Bone–Implant Contact in Three Short Dental Implant Models with Varying Osseointegration Percentages

Objectives: Dental implants have become a cornerstone of restorative dentistry, providing a long-lasting method for tooth replacement. The degree of osseointegration has a significant effect on biomechanical stability at the bone–implant contact (BIC), determining the continued efficacy of these implants. However, the exact consequences of changing osseointegration levels on different implant designs, especially in bones with variable densities, are not well known. Methods: This study used 3D finite element analysis (FEA) to look at the biomechanical performance of three short dental implants: BioMet 3iT3, Straumann® Standard Plus Short-Regular Neck (SPS-RN), and Straumann® Standard Plus Short-Wide Neck (SPS-WN). This paper tested the implants at four stages of osseointegration: 25%, 50%, 75%, and 100% in both high-density (bone type III) and low-density (bone type IV) cancellous bone. It also created and examined realistic CAD models under static occlusal loading conditions to assess stress distribution and major strains at the bone–implant contact. Results: The study discovered that as osseointegration increases, von Mises stress and principal strains go down significantly for all implant types. The SPS-WN implant had the lowest strain values, especially for bone with low density. These reductions demonstrate increased mechanical stability as the bone–implant interface becomes more capable of dispersing mechanical stresses, minimizing the potential for localized deformation and bone resorption. Conclusions: The results highlight the importance of achieving optimum osseointegration to reduce mechanical stress and increase the lifespan of dental implants. The SPS-WN type implant performed better in biomechanical tests than the others, especially when bone conditions were not ideal. This makes it a great choice for clinical applications that need long-term implant success.

DPF-YOLOv8: Dual Path Feature Fusion Network for Traffic Sign Detection in Hazy Weather

Traffic sign detection plays an integral role in intelligent driving systems. It was found that in real driving scenarios, traffic signs were easily obscured by haze leading to traffic sign detection inaccuracy in assisted driving systems. Therefore, we designed a traffic sign detection model for hazy weather that can effectively help drivers to recognize road signs and reduce the incidence of traffic accidents. A high-precision traffic sign detection network has been designed to address the problem of decreased model recognition performance caused by external factors such as small size of traffic signs and haze obstruction in real-world scenarios. First, the default YOLOv8 was found to have low model detection accuracy in hazy weather occlusion conditions through experimental studies. Therefore, a deeper lightweight and efficient multi-branch CSP (Cross Stage Partial) module was introduced. Second, a dual path feature fusion network was designed to address the problem of insufficient feature fusion due to the small size of traffic signs. Finally, in order to be able to better simulate the real haze weather scene, we added fog to the raw data to enrich the data samples. This was verified through experiments on a public Chinese traffic sign detection dataset after fogging treatment, compared to the default YOLOv8 model. The improved DPF-YOLOv8 algorithm achieved 2.1% and 2.2% improvement in mAP@0.5 and mAP@0.5:0.95 performance metrics to 65.0% and 47.4%, respectively.

Whole-exome sequencing reveals the genetic causes and modifiers of moyamoya syndrome

Moyamoya vasculopathy secondary to various genetic disorders is classified as moyamoya syndrome (MMS). Recent studies indicate MMS occurs due to a combination of genetic modifiers and causative mutations for the primary genetic disorders. We performed whole-exome sequencing (WES) in 13 patients with various genetic disorders who developed MMS. WES successfully revealed the genetic diagnoses of neurofibromatosis type 1 (NF-1), Down syndrome, multisystemic smooth muscle dysfunction syndrome, Noonan syndrome, and alpha thalassemia. The previously reported modifier genes, RNF213 and MRVI1, were confirmed in the NF-1 and Down syndrome cases. Further analysis revealed rare hypomorphic variants in the causative genes of the primary disorders underlying MMS, such as Alagille syndrome and Rasopathies, conferred susceptibility to MMS. Genes involved in the development of pulmonary arterial hypertension (PAH), such as ABCC8 and BMPR2, were also identified as potential modifiers. The rare variants in the MMS and PAH genes were significantly enriched in the eight Japanese patients with MMS compared with the 104 Japanese individuals from the 1000 Genomes Project. Disease genes associated with the arterial occlusive conditions represented by those of Rasopathies and PAH may provide novel diagnostic markers and future therapeutic targets for MMS as well as moyamoya disease with an unknown cause.

Defining ideal middle cerebral artery bifurcation aneurysm size for Woven EndoBridge embolization.

The Woven EndoBridge (WEB) device was approved to treat wide-necked bifurcation aneurysms. The device is designed as an intrasaccular flow disruptor covering aneurysm widths up to 10 mm. Although prior studies combined all aneurysm sizes, it is known that aneurysms behave differently in response to endovascular treatment based on their size. Therefore, the authors' objective was to identify ideal middle cerebral artery (MCA) aneurysm width and neck sizes most suitable for WEB treatment. The WorldWideWEB consortium is a large multicenter retrospective database that analyzes intracranial aneurysms treated with the WEB device. In this study, all unruptured MCA bifurcation aneurysms with available measurements were included. Cutoff values based on aneurysm width and neck in relation to aneurysm occlusion status were measured using the receiver operating characteristic (ROC) curve. Propensity score matching (PSM) was then used to compare treatment outcomes between aneurysms smaller and larger than the cutoff value for both width and neck size. The ideal cutoff values for MCA bifurcation aneurysm width and neck were 6.1 mm and 4.6 mm, respectively. On PSM, 87 matched pairs were compared based on width size (≤ 6.1 mm and > 6.1 mm), and 77 matched pairs were compared based on neck size (≤ 4.6 mm and > 4.6 mm). There was a significant difference in adequate aneurysm occlusion between aneurysms smaller and larger than those cutoff values for both widths (93% vs 76%, p = 0.0017) and neck sizes (90% vs 70%, p = 0.0026). The retreatment rate was also significantly higher for larger aneurysms in both parameters. This study shows that MCA bifurcation aneurysms ≤ 6.1 mm in width and ≤ 4.6 mm in neck size are significantly better candidates for WEB treatment, leading to improved occlusion status and reduced retreatment rate, which are important considerations when using WEB devices.

Advanced 3D Face Reconstruction from Single 2D Images Using Enhanced Adversarial Neural Networks and Graph Neural Networks.

This paper presents a novel framework for 3D face reconstruction from single 2D images and addresses critical limitations in existing methods. Our approach integrates modified adversarial neural networks with graph neural networks to achieve state-of-the-art performance. Key innovations include (1) a generator architecture based on Graph Convolutional Networks (GCNs) with a novel loss function and identity blocks, mitigating mode collapse and instability; (2) the integration of facial landmarks and a non-parametric efficient-net decoder for enhanced feature capture; and (3) a lightweight GCN-based discriminator for improved accuracy and stability. Evaluated on the 300W-LP and AFLW2000-3D datasets, our method outperforms existing approaches, reducing Chamfer Distance by 62.7% and Earth Mover's Distance by 57.1% on 300W-LP. Moreover, our framework demonstrates superior robustness to variations in head positioning, occlusion, noise, and lighting conditions while achieving significantly faster processing times.

Bud-YOLO: A Real-Time Accurate Detection Method of Cotton Top Buds in Cotton Fields

Cotton topping plays a crucial and indispensable role in controlling excessive growth and enhancing cotton production. This study aims to improve the operational efficiency and accuracy of cotton topping robots through a real-time and accurate cotton top bud detection algorithm tailored for field operation scenarios. We propose a lightweight structure based on YOLOv8n, replacing the C2f module with the Cross-Stage Partial Networks and Partial Convolution (CSPPC) module to minimize redundant computations and memory access. The network’s neck employs an Efficient Reparameterized Generalized-FPN (Efficient RepGFPN) to achieve high-precision detection without substantially increasing computational cost. Additionally, the loss calculation of the optimized prediction frame was addressed with the Inner CIoU loss function, thereby enhancing the precision of the model’s prediction box. Comparison experiments indicate that the Bud-YOLO model is highly effective for detecting cotton top buds, with an AP50 of 99.2%. This performance surpasses that of other YOLO variants, such as YOLOv5s and YOLOv10n, as well as the conventional Faster R-CNN model. Moreover, the Bud-YOLO model exhibits robust performance across various angles, occlusion conditions, and bud morphologies. This study offers technical insights to support the migration and deployment of the model on cotton topping machinery.