Success Rate Research Articles

Unveiling the Threat: Investigating Distributed and Centralized Backdoor Attacks in Federated Graph Neural Networks

Graph neural networks (GNNs) have gained significant popularity as powerful deep learning methods for processing graph data. However, centralized GNNs face challenges in data-sensitive scenarios due to privacy concerns and regulatory restrictions. Federated learning has emerged as a promising technology that enables collaborative training of a shared global model while preserving privacy. Although federated learning has been applied to train GNNs, no research focuses on the robustness of Federated GNNs against backdoor attacks. This article bridges this research gap by investigating two types of backdoor attacks in Federated GNNs: centralized backdoor attack (CBA) and distributed backdoor attack (DBA). Through extensive experiments, we demonstrate that DBA exhibits a higher success rate than CBA across various scenarios. To further explore the characteristics of these backdoor attacks in Federated GNNs, we evaluate their performance under different scenarios, including varying numbers of clients, trigger sizes, poisoning intensities, and trigger densities. Additionally, we explore the resilience of DBA and CBA against two defense mechanisms. Our findings reveal that both defenses cannot eliminate DBA and CBA without affecting the original task. This highlights the necessity of developing tailored defenses to mitigate the novel threat of backdoor attacks in Federated GNNs.

Development of the Children’s ACuTe UltraSound (CACTUS) point-of-care ultrasound (POCUS)-accredited training in the UK: a descriptive study

Point-of-care ultrasound (POCUS) is an established, evidence-supported tool that can be used in neonatal and paediatric medicine, offering clinicians immediate diagnostic insights, assessment of interventions and improved safety profiles and...

Heatmap creation with YOLO-Deep SORT system customized for in-store customer behavior analysis

Due to the limitations of the hardware system, analysis of retail stores has caused problems such as excessive workload, incomplete analysis, slow analysis speed, difficult data collection, non-real-time data collection, passenger flow statistics, and density analysis. However, heatmaps are a viable solution to these problems and provide adaptable and effective analysis. In this paper, we propose to use the deep sequence tracking algorithm together with the YOLO object recognition algorithm to create heatmap visualizations. We will present key innovations of our customized YOLO-Deep SORT system to solve some fundamental problems in in-store customer behavior analysis. These innovations include our use of footpad targeting to make bounding boxes more precise and less noisy. Finally, we made a comprehensive evaluation and comparison to determine the success rate of our system and found that the success rate was higher than the systems we compared in the literature. The results show that our heatmap visualization enables accurate, timely, and detailed analysis.

Evaluation of post-surgical complications between "Rendez-vous" technique vs. standard care in patients with choledocholithiasis

"Rendez-vous" (RV) technique is a mixed-technique which uses both laparoscopic and endoscopic skills; however, the evidence is contradictory regarding the implementation of this technique or the 2-step sequential technique (endoscopic retrograde cholangiopancreatography [ERCP] followed by laparoscopic cholecystectomy [LC]) in the management of cholecysto-choledocholitiasis. To estimate the association between the implementation of RV technique and the presence of post-surgical complications as primary outcome, using as comparator the 2-step sequential technique. An observational, analytical, retrospective study was conducted, using as exposed cohort the medical records from patients with a diagnosis of cholelithiasis, cholecystitis, or mild biliary pancreatitis. The exposed cohort underwent RV technique, while the unexposed cohort were those which underwent two step technique. There was a lower post-surgical complication rate in the RV group (0%) compared with the 10.1% (p = 0.3617) in the control group. Also, RV technique showed a lesser hospitalization time (p = 0.0377) and a lesser post-surgical hospitalization time (p < 0.0001). RV technique is superior when compared with the 2-step sequential technique (ERCP followed by LC), based on a better surgical success rate, a fewer complications rate and less hospitalization time.

Cube-Evo: A Query-Efficient Black-Box Attack on Video Classification System

The current progressive research in the domain of black-box adversarial attack enhances the reliability of deep neural network (DNN)-based video systems. Recent works mainly carry out black-box adversarial attacks on video systems by query-based parameter dimension reduction. However, the additional temporal dimension of video data leads to massive query consumption and low attack success rate. In this article, we embark on our efforts to design an effective adversarial attack on popular video classification systems. We deeply root the observations that the DNN-based systems are sensitive to adversarial perturbations with high frequency and reconstructed shape. Specifically, we propose a systematic attack pipeline Cube-Evo, aiming to reduce the search space dimension and obtain the effective adversarial perturbation via the optimal parameter group updating. We evaluate the proposed attack pipeline on two popular datasets: UCF101 and JESTER. Our attack pipeline reduces query consumption and achieves a high success rate on various DNN-based video classification systems. Compared with the state-of-the-art method Geo-Trap-Att, our pipeline averagely reduces 1.6× query consumption in untargeted attacks and 2.9× in targeted attacks. Besides, Cube-Evo improves 13% attack success rate on average, achieving new state-of-the-art results over diverse video classification systems.

Parathyroidectomy: still the best choice for the management of severe secondary hyperparathyroidism.

Management of secondary hyperparathyroidism (SHPT) is a challenging endeavor with several factors contruibuting to treatment failure. Calcimimetic therapy has revolutionized the management of SHPT, leading to changes in indications and appropriate timing of parathyroidectomy (PTX) around the world. We compared response rates to clinical vs. surgical approaches to SHPT in patients on maintenance dialysis (CKD 5D) and in kidney transplant patients (Ktx). A retrospective analysis of the one-year follow-up findings was carried out. CKD 5D patients were divided into 3 groups according to treatment strategy: parathyroidectomy, clinical management without cinacalcet (named standard - STD) and with cinacalcet (STD + CIN). Ktx patients were divided into 3 groups: PTX, CIN (cinacalcet use), and observation (OBS). In CKD 5D we found a significant parathormone (PTH) decrease in all groups. Despite all groups had a higher PTH at baseline, we identified a more pronounced reduction in the PTX group. Regarding severe SHPT, the difference among groups was evidently wider: 31%, 14% and 80% of STD, STD + CIN, and PTX groups reached adequate PTH levels, respectively (p<0.0001). Concerning the Ktx population, although the difference was not so impressive, a higher rate of success in the PTX group was also observed. PTX still seems to be the best treatment choice for SHPT, especially in patients with prolonged diseases in unresourceful scenarios.

A retrospective comparison of endoscopic nasobiliary drainage (ENBD) and endoscopic retrograde biliary drainage (ERBD) in the treatment of acute cholangitis

Endoscopic nasobiliary drainage (ENBD) and endoscopic retrograde biliary drainage (ERBD) can rapidly alleviate obstruction, facilitate bile drainage, and reduce pressure in the bile duct. In this study, to evaluate the effect of plastic biliary stent on the recurrence rate of choledocholithiasis in patients with actual cholangitis after endoscopic retrograde cholangioangiography (ERCP), we a comparative analysis of the efficacy and safety between ENBD and ERBD in inflammatory remission rate, complication rate, and success rate of secondary or multiple endoscopic lithotomy. A total of 347 patients undergoing ENBD and ERBD were enrolled in this retrospective study. Preoperative assessments including blood tests, electrocardiography, abdominal ultrasonography, CT scans, MRI, or MRCP were conducted on all participants. Additionally, serum CRP and amylase levels were measured using a magnetic nanobead-based immunoassay. We found that The success rate of stone removal in the ENBD group was higher than that in the ERBD group (90.8% vs. 72.7%) (P &lt;0.0001). The remission rate of inflammation in the ERBD group (100.0%) was higher than that in the ENBD group (98.7%) (P &gt;0.05). No significant difference was observed between the ERBD group and the ENBD group in preoperative and postoperative inflammatory indexes (total bilirubin, direct bilirubin, white blood cells, neutrophils, CRP, and serum amylase). Moreover, there were significant differences in post-ERCP pancreatitis between the ERBD group and ENBD group. These results suggest that both ENBD and ERBD can effectively improve clinical efficacy. Moreover, magnetic nanobeads have good application in serum index detection to increase detection efficiency.

Efficacy and safety of a neurointerventional operation robotic assistance system in cerebral angiography

BackgroundAt present, neurointerventional surgery requires angiographers to perform operations in the digital subtraction angiography (DSA) room. Ionising radiation and chronic joint damage are still unavoidable for angiographers. Therefore, we researched...

Laser lead extraction of very old leads. Insights from the GermAn Laser Lead Extraction RegistrY (GALLERY)

Abstract Background Transvenous lead extraction (TLE) has become one of the most crucial therapeutic pillars in modern CIED management. Very old pacemaker and defibrillator leads are known to be risk factors for procedural complications in patients undergoing TLE. Aim &amp; Methods Aim of this study was to analyze patients with very old leads undergoing TLE. We conducted a retrospective analysis of all patients in the GALLERY with at least one lead aged ≥10 years and compared them to a cohort of patients with leads aged &amp;lt;10 years, in terms of baseline- and lead characteristics, as well as procedural outcomes and potential predictors of adverse events. Results We identified 858 patients with at least one lead aged ≥10 years (Group A) and 1666 patients with leads aged &amp;lt;10 years (Group B). Patients in Group A were older (70.4±13.7 vs. 66.7±13.7; p&amp;lt;0.001) and were less likely to exhibit highly reduced LVEF (18.5 vs. 30.0%, p&amp;lt;0.001) or coronary artery disease (36.1 vs. 46.4%, p&amp;lt;0.001), compared to Group B. The mean numbers of leads per patient were 2.5±1.1 and 2.4±1.0, respectively (p&amp;lt;0.001) with median ages of the oldest indwelling lead of 161.5 [137-206.75] and 72.0 [48-94] months (p&amp;lt;0.001). Patients in Group A had higher proportions of right-sided (47.7 vs. 25.8%, p&amp;lt;0.001) and abandoned leads (41.3 vs. 23.4%, p&amp;lt;0.001). Median procedural time was longer in Group A with 93.5 [63-145] vs. 75 [52-116] minutes (p&amp;lt;0.001) with an increased necessity for additional tools for extraction (12.6 vs. 3.4%, p&amp;lt;0.001). Overall (5.5 vs. 3.6%, p=0.03) and major complication rates (3.2 vs. 1.4%, p=0.005) were both higher in Group A. Patients in Group A were less likely to achieve complete procedural success rate (84.3 vs. 95.1%, p&amp;lt;0.001) and had a higher rate for procedural failure (3.5 vs. 1.5%, p=0.002). There was no difference in procedure-related &amp;lt;(0.7 vs. 0.5%, p=0.592) or all-cause mortality (3.5 vs. 3.2%, p=0.711) between groups. Multivariate analysis in Group A revealed the presence of abandoned leads (OR: 1.83, 95% CI: 1.0-3.4, p=0.049), female sex (OR: 2.15, 95% CI: 1.2-3.9, p=0.013) and systemic infection (OR:1.89, 95%CI: 1.0-3.5, p=0.039) as predictors for procedural complications. The use of additional extraction tools (OR:4.4, 95%CI:2.0-9.7, p&amp;lt;0.001) and a BMI≥35 kg/m2 (OR:4.4, 95%CI:1.4-14.0, p=0.012) were identified as risk factors for procedural failure. Procedural complications (OR:5.2, 95%CI:2.0-13.7, p&amp;lt;0.001) and systemic infection (OR:18.7, 95%CI: 5.6-62.9, p&amp;gt;0.001) were highly predictive for all-cause mortality. Conclusion Patients with leads aged ≥10 years undergoing TLE suffer from more procedural complications, more often require additional extraction tools and have lower rates for complete procedural success, but do not show any differences in mortality, when compared to patients with leads aged &amp;lt;10 years. The necessity of additional tools for extraction and a BMI≥35 kg/m2 were identified as predictors for procedural failure.

Can critical care transport be safely reduced in children intubated during emergency management of status epilepticus in the United Kingdom: a national audit with case–control analysis

ObjectiveThis study describes the baseline clinical characteristics, predictors of successful extubation at referring hospitals and short-term outcomes of children intubated for status epilepticus and referred to United Kingdom (UK) paediatric...

TomFi: Small Object Tracking Using Commodity WiFi

Rodent infestation has always been one of the most severe threats to humans, which to solve consumes massive manpower and resources. People usually use traps or poisons to treat rat infestation. Such passive countermeasures are inefficient. Damage often occurs when the rats are finally trapped or killed, not to mention the potential risk to injure humans or cause pollution. In this article, we propose a WiFi-based active small object tracking system named TomFi . The core components of TomFi include several deep learning techniques that bridge rat locations/motions and WiFi signal variations (represented by the channel state information). TomFi first employs a detection model to detect the appearance of the rat and then localize it based on a two-branch localization model. Through the design of the two-branch localization model, to our best knowledge, we are the first to solve the information loss and distortion problem. We conducted extensive experiments in both the laboratory environment and real-world kitchen scenarios. The results show that TomFi can achieve a detection success rate of 99%+ and a centimeter-level localization accuracy in real time.

Point Cloud Attacks in Graph Spectral Domain: When 3D Geometry Meets Graph Signal Processing.

With the increasing attention in various 3D safety-critical applications, point cloud learning models have been shown to be vulnerable to adversarial attacks. Although existing 3D attack methods achieve high success rates, they delve into the data space with point-wise perturbation, which may neglect the geometric characteristics. Instead, we propose point cloud attacks from a new perspective-the graph spectral domain attack, aiming to perturb graph transform coefficients in the spectral domain that correspond to varying certain geometric structures. Specifically, leveraging on graph signal processing, we first adaptively transform the coordinates of points onto the spectral domain via graph Fourier transform (GFT) for compact representation. Then, we analyze the influence of different spectral bands on the geometric structure, based on which we propose to perturb the GFT coefficients via a learnable graph spectral filter. Considering the low-frequency components mainly contribute to the rough shape of the 3D object, we further introduce a low-frequency constraint to limit perturbations within imperceptible high-frequency components. Finally, the adversarial point cloud is generated by transforming the perturbed spectral representation back to the data domain via the inverse GFT. Experimental results demonstrate the effectiveness of the proposed attack in terms of both the imperceptibility and attack success rates.

Intravitreal bevacizumab improves trabeculectomy survival at 12 months: the bevacizumab in trabeculectomy study—a randomised clinical trial

AimsTo evaluate the effect of an intraoperative dose of intravitreal bevacizumab (Avastin) on surgical success following trabeculectomy with mitomycin-C (MMC) over 12 months.MethodsA single centre, parallel, double-blinded randomised, placebo-controlled trial...

A Contemporary Review on the Critical Role of Nonsteroidal Anti-inflammatory Agents in Colorectal Cancer Therapy.

Nonsteroidal Anti-Inflammatory Drugs (NSAIDs) are widely recognized as effective pain relievers and function by inhibiting the cyclooxygenase enzyme (COXs). Moreover, they have been found to participate in various cellular processes through different signaling pathways, such as WNT, MAPK, NF-κB, and PI3K/AKT/mTOR. This makes them potential candidates for chemoprevention of several malignancies, particularly colorectal cancer (CRC). However, the use of NSAIDs in cancer prevention and treatment is a complex issue due to their adverse effects and gastrointestinal toxicity. Therefore, it is crucial to explore combination therapies that can minimize side effects while maximizing synergistic effects with other agents and to evaluate the success rate of such approaches in both pre-clinical and clinical studies. In this review, we aim to provide an overview of the effects of NSAIDs in the prevention and treatment of CRC. We will focus on elucidating the possible mechanisms of action of these drugs, the signaling pathways involved in CRC, and the potential synergistic effects when combined with other therapeutic agents.

Comparison of Surgical and Colonoscopy Tissue to Establish Colorectal Patient-derived Organoids.

Patient-derived organoids (PDOs) are ex vivo models that retain the functions and characteristics of individualized source tissues, including a simulated tumor microenvironment. However, the potential impact of undiscovered differences between tissue sources on PDO growth and progression remains unclear. This study aimed to compare the growth and condition of PDO models originating from surgical resection and colonoscopy and to provide practical insights for PDO studies. Tissue samples and relevant patient clinical information were collected to establish organoid models. PDOs were derived from both surgical and colonoscopy tissues. The growth of the organoids, including their state, size, and success rate of establishment, was recorded and analyzed. The activity of the organoids at the end stage of growth was detected using calcein-AM fluorescence staining. The results showed that the early growth phase of 2/3 colonoscopy-derived organoids was faster compared to surgical PDOs, with a growth difference observed within 11-13 days of establishment. However, colonoscopy-derived organoids exhibited a diminished growth trend after this time. There were no significant differences observed in the terminal area and quantity between the two types of tissue-derived organoids. Immunofluorescence assays of the PDOs revealed that the surgical PDOs possessed a denser cell mass with relatively higher viability than colonoscopy-derived PDOs. In the establishment of colorectal patient-derived organoids, surgically derived organoids require a slightly longer establishment period, while colonoscopy-derived organoids should be passaged prior to growth inhibition to preserve organoid viability.

Quantization Backdoors to Deep Learning Commercial Frameworks

Due to their low latency and high privacy preservation, there is currently a burgeoning demand for deploying deep learning (DL) models on ubiquitous edge Internet of Things (IoT) devices. However, DL models are often large in size and require large-scale computation, which prevents them from being placed directly onto IoT devices, where resources are constrained, and 32-bit floating-point (float-32) operations are unavailable. Commercial framework (i.e., a set of toolkits) empowered model quantization is a pragmatic solution that enables DL deployment on mobile devices and embedded systems by effortlessly post-quantizing a large high-precision model (e.g., float-32) into a small low-precision model (e.g., int-8) while retaining the model inference accuracy. However, their usability might be threatened by security vulnerabilities. This work reveals that standard quantization toolkits can be abused to activate a backdoor. We demonstrate that a full-precision backdoored model which does not have any backdoor effect in the presence of a trigger—as the backdoor is dormant—can be activated by (i) TensorFlow-Lite (TFLite) quantization, the only <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">product-ready</i> quantization framework to date, and (ii) the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">beta released</i> PyTorch Mobile framework. In our experiments, we employ three popular model architectures (VGG16, ResNet18, and ResNet50), and train each across three popular datasets: MNIST, CIFAR10 and GTSRB. We ascertain that all trained float-32 backdoored models exhibit no backdoor effect <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">even in the presence of trigger inputs</i> . Particularly, four influential backdoor defenses are evaluated, and they fail to identify a backdoor in the float-32 models. When each of the float-32 models is converted into an int-8 format model through the standard TFLite or PyTorch Mobile framework's post-training quantization, the backdoor is activated in the quantized model, which shows a stable attack success rate close to 100% upon inputs with the trigger, while it usually behaves upon non-trigger inputs. This work highlights that a stealthy security threat occurs when an end-user utilizes the on-device post-training model quantization frameworks, informing security researchers of a cross-platform overhaul of DL models post-quantization even if these models pass security-aware front-end backdoor inspections. Significantly, we have identified Gaussian noise injection into the malicious full-precision model as an easy-to-use preventative defense against the PQ backdoor. The attack source code is released at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/quantization-backdoor</uri> .

Clinical, Radiographic, and Histologic Outcomes of Regenerative Endodontic Treatment in Human Immature Teeth Using Different Biological Scaffolds: A Systematic Review and Meta-analysis.

Biological scaffolds such as blood clot (BC), platelet-rich plasma (PRP), platelet- rich fibrin (PRF), and platelet pellet (PP) are used in regenerative endodontic treatments (RETs). To systematically and quantitatively evaluate clinical, radiographic, and histologic outcomes of RET studies using different biological scaffolds. MEDLINE, Scopus, Cochrane library, and Embase were searched to identify studies on RET procedures with any scaffold type performed on immature non-vital human teeth, employing any type of biological scaffold. Clinical, radiographic, and histologic outcomes were extracted. Cochrane collaboration risk of bias tool and Newcastle-Ottawa scale were used for quality assessment. Random and fixed model meta-analysis was carried out with 95% confidence interval. Thirty-two studies were included in the qualitative analysis from the primarily retrieved 1895 studies. Only one study had high risk of bias and 71.8% of the studies had high quality. None of the studies reported any histologic findings. Thirty studies were included in meta-analysis. Clinical success rate of RET using either BC, PRP, or PRF was >99%. Furthermore, 32%, 23%, and 27% of BC, PRP, and PRF cases regained vitality, respectively. Periapical healing was seen in 67%, 75%, and 100% of BC, PRP, and PRF cases, respectively. There was no statistical difference between BC, PRP, or PRF regarding clinical success or any radiographic outcomes. There was no significant difference between BC, PRP, and PRF in terms of clinical and radiographic outcomes. When it is difficult or dangerous to induce bleeding in root canals, PRP and PRF may be employed instead.

Success rates and failures of fixed and removable space maintainers after the premature loss of primary molars.

The evidence base for the use of space maintainers is relatively sparce despite being used for decades after the premature loss of primary molars. This study aims to increase the dental evidence base via investigating retrospectively the success rates of prefabricated fixed and removable space maintainers inserted from 2019 to 2021 and followed up until February 2023 at a specialized university clinic and to identify reasons for any reported minor and major failure. The authors hypothesized that there is no significant difference in failure rates between fixed and removable space maintainers inserted after the premature loss of a single primary molar per quadrant. Patients' digital records were searched yielding 645 space maintainers. After the application of inclusion criteria, 157 (67%) fixed prefabricated space maintainers in 112 children and 77 (33%) removable space maintainers in 61 children were analyzed for an average of 18.4 ± 9.5 months. Kaplan-Meier survival analysis with Mantel-Cox statistics showed an overall cumulative survival time of 31.6 months (SE = 1.15, 95% CI = 29.4 to 33.9). Major failure occurred significantly more in removable maintainers (n = 40/67, 59.7%), mostly due to loss of the appliance, compared to fixed space maintainers (n = 27/67, 40.3%; P < .001). The present study indicates that space maintainers were mainly placed in young children with high caries experience, where treatment was mostly possible using advanced behavior management. Fixed space maintainers had a significantly lower failure rate than their removable counterpart. However, both require continual repairs, preservation, or even replacement till the eruption of the permanent tooth.

Return to Sports in Patients Aged 50 Years or Younger After Robotic-Assisted Patellofemoral Arthroplasty: A 10-Year Experience Reporting High Clinical Benefits and High Patient Satisfaction With Return to an Active Lifestyle

Background: Limited data are available on return to sports and patient psychometric ratings of success after patellofemoral arthroplasty (PFA) in younger patients with high expectations to return to an active lifestyle. Purpose/Hypothesis: The purpose of this article was to determine the role of PFA and its success in meeting patient expectations regarding the return to low-impact recreational sports and an active lifestyle in younger, active patients. It was hypothesized that PFA would allow younger patients to return to low-impact sports and an active lifestyle and achieve high patient psychometric ratings. Study Design: Case series; Level of evidence, 4. Methods: In this 10-year prospective study (2009-2018), robotic-assisted PFA was performed on 44 patients (32 women and 12 men; n = 51 consecutive knees), with a mean age of 37.2 years (range, 21-50 years). The follow-up rate was 98%, with a mean of 5.3 years (range, 2-9.3 years). Primary clinical outcomes were as follows: the validated Cincinnati Knee Rating System sports activity and symptom rating scales, patient psychometric ratings of the substantial clinical benefit (SCB), and the Patient Acceptable Symptom State (PASS). Secondary outcomes were the Cincinnati Knee Rating System occupational rating, visual analog pain scale, and the 12-Item Short Form Health Survey. Survivorship was defined by conversion to total knee replacement (TKR). Results: Before PFA, 78% of patients (35/45 knees) were symptomatic and unable to perform recreational sports, with only 20% of patients (9/45 knees) performing some low-impact sports. After PFA, 80% of patients (36/45 knees) were able to perform low-impact sports, and 7% (3/45 knees) performed jumping-pivoting sports (P < .001). The SCB scored by the patient showed 87% of knees as good, very good, or normal. On the PASS analysis, 89% of patients (95% CI, 76%-96%) were “pleased,” and 93% (95% CI, 82%-99%) would undergo surgery again. There were clinically relevant improvements in symptoms of pain, swelling, and giving way (P = .0001). Preoperatively, 91% of knees had moderate to severe pain with activities of daily living, and only 11% of knees had pain at the follow-up. Five of the 50 knees (10%) underwent TKR conversion with one patient lost to follow-up. Conclusion: PFA resulted in a high return of patients to low-impact sports with high SCB and PASS psychometric ratings. The robotic-assisted 3-dimensional preoperative planning allowed precise intraoperative trochlear implant alignment in knees with severe trochlear dysplasia. PFA is recommended as an alternative treatment in younger patients with end-stage symptomatic patellofemoral arthritis. Registration: NCT02738476 (ClinicalTrials.gov identifier).

Machine and Deep Learning Implementations for Heritage Building Information Modelling: A Critical Review of Theoretical and Applied Research

Research domain and Problem: HBIM modelling from point cloud data has become a crucial research topic in the last decade since it is potentially considered as the central data model paving the way for the digital heritage practice beyond digitization. Reality Capture technologies such as terrestrial laser scanning, drone-mounted LiDAR sensors and photogrammetry enable the reality capture with a sub-millimetre accurate point cloud file that can be used as a reference file for Heritage Building Information Modelling (HBIM). However, HBIM modelling from the point cloud data of heritage buildings is mainly manual, error-prone, and time-consuming. Furthermore, image processing techniques are insufficient for classification and segmentation of point cloud data to speed up and enhance the current workflow for HBIM modelling. Due to the challenges and bottlenecks in the scan-to-HBIM process, which is commonly criticized as complex with its bespoke requirements, semantic segmentation of point clouds is gaining popularity in the literature. Research Aim and Methodology: Therefore, this paper aims to provide a thorough critical review of Machine Learning and Deep Learning methods for point cloud segmentation, classification, and BIM geometry automation for cultural heritage case study applications. Research findings: This paper files the challenges of HBIM practice and the opportunities for semantic point cloud segmentation found across academic literature in the last decade. Beyond definitions and basic occurrence statistics, this paper discusses the success rates and implementation challenges of machine and deep learning classification methods. Research value and contribution: This paper provides a holistic review of point cloud segmentation and its potential for further development and application in the Cultural Heritage sector. The critical analysis provides insight into the current state-of-the-art methods and advises on their suitability for HBIM projects. The review has identified highly original threads of research, which hold the potential to significantly influence practice and further applied research.