Fully Self-Supervised Out-of-Domain Few-Shot Learning with Masked Autoencoders.

Few-shot learning aims to identify unseen classes with limited labelled data. Recent few-shot learning techniques have shown success in generalizing to unseen classes; however, the performance of these techniques has also been shown to degrade when tested on an out-of-domain setting. Previous work, additionally, has also demonstrated increasing reliance on supervised finetuning in an off-line or online capacity. This paper proposes a novel, fully self-supervised few-shot learning technique (FSS) that utilizes a vision transformer and masked autoencoder. The proposed technique can generalize to out-of-domain classes by finetuning the model in a fully self-supervised method for each episode. We evaluate the proposed technique using three datasets (all out-of-domain). As such, our results show that FSS has an accuracy gain of 1.05%, 0.12%, and 1.28% on the ISIC, EuroSat, and BCCD datasets, respectively, without the use of supervised training.

Masked Embedding Modeling With Rapid Domain Adjustment for Few-Shot Image Classification.

In few-shot classification, performing well on a testing dataset is a challenging task due to the restricted amount of labelled data available and the unknown distribution. Many previously proposed techniques rely on prototypical representations of the support set in order to classify a query set. Although this approach works well with a large, in-domain support set, accuracy suffers when transitioning to an out-of-domain setting, especially when using small support sets. To address out-of-domain performance degradation with small support sets, we propose Masked Embedding Modeling for Few-Shot Learning (MEM-FS), a novel, self-supervised, generative technique that reinforces few-shot-classification accuracy for a prototypical backbone model. MEM-FS leverages the data completion capabilities of a masked autoencoder to expand a given embedded support set. To further increase out-of-domain performance, we also introduce Rapid Domain Adjustment (RDA), a novel, self-supervised process for quickly conditioning MEM-FS to a new domain. We show that masked support embeddings generated by MEM-FS+RDA can significantly improve backbone performance on both out-of-domain and in-domain datasets. Our experiments demonstrate that applying the proposed technique to an inductive classifier achieves state-of-the-art performance on mini-imagenet, the CVPR L2ID Classification Challenge, and a newly proposed dataset, IKEA-FS. We provide code for this work at https://github.com/Brikwerk/MEM-FS.

Automated human cell classification in sparse datasets using few-shot learning.

Classifying and analyzing human cells is a lengthy procedure, often involving a trained professional. In an attempt to expedite this process, an active area of research involves automating cell classification through use of deep learning-based techniques. In practice, a large amount of data is required to accurately train these deep learning models. However, due to the sparse human cell datasets currently available, the performance of these models is typically low. This study investigates the feasibility of using few-shot learning-based techniques to mitigate the data requirements for accurate training. The study is comprised of three parts: First, current state-of-the-art few-shot learning techniques are evaluated on human cell classification. The selected techniques are trained on a non-medical dataset and then tested on two out-of-domain, human cell datasets. The results indicate that, overall, the test accuracy of state-of-the-art techniques decreased by at least 30% when transitioning from a non-medical dataset to a medical dataset. Reptile and EPNet were the top performing techniques tested on the BCCD dataset and HEp-2 dataset respectively. Second, this study evaluates the potential benefits, if any, to varying the backbone architecture and training schemes in current state-of-the-art few-shot learning techniques when used in human cell classification. To this end, the best technique identified in the first part of this study, EPNet, is used for experimentation. In particular, the study used 6 different network backbones, 5 data augmentation methodologies, and 2 model training schemes. Even with these additions, the overall test accuracy of EPNet decreased from 88.66% on non-medical datasets to 44.13% at best on the medical datasets. Third, this study presents future directions for using few-shot learning in human cell classification. In general, few-shot learning in its current state performs poorly on human cell classification. The study proves that attempts to modify existing network architectures are not effective and concludes that future research effort should be focused on improving robustness towards out-of-domain testing using optimization-based or self-supervised few-shot learning techniques.

Improvement of Coronary Calcium Scores After Bariatric Surgery in People with Severe Obesity.

BACKGROUND: Obesity is an independent risk factor for cardiovascular diseases. Coronary artery calcium (CAC) is a direct measure of coronary atherosclerosis. The study investigated the effect of bariatric surgery on CAC scores in people with severe obesity subjected to laparoscopic sleeve gastrectomy (LSG). METHODS: This prospective study included 129 people with severe obesity in two groups; the LSG group (n=74) subjected to surgery and the diet group (n=55), managed by a diet regimen and lifestyle modification. Cardiovascular risk was assessed by Framingham risk score (FRS) and coronary calcium score (CCS) measured by computed tomography initially and after 3 years. RESULTS: The two groups had a comparable CAD risk before treatment according to FRS or CCS. After treatment, CCS improved significantly in the LSG group (p=0.008) but not in the diet group (p=0.149). There was no correlation between FRS and CCS (r=0.005, p=0.952). Treatment resulted in significant weight reduction and improved fasting blood glucose and lipid profile in the two groups. The change of weight, blood glucose, and HDL, and remission of diabetes mellitus (DM) were significantly higher in the LSG group compared to the diet group. CONCLUSION: LSG may reduce the risk of developing future cardiovascular comorbidities evidenced by reducing CAC scores. Significant weight reduction and improvement of cardiovascular risk factors may recommend LSG as a cardioprotective procedure in people with severe obesity.

EFFICACY OF HEMOSTATIC AGENTS IN ENDODONTIC SURGERY: A SYSTEMATIC REVIEW AND NETWORK META-ANALYSIS.

OBJECTIVE: Adequate hemostasis is a critical step in endodontic surgery. It facilitates the procedure and affects the success and prognosis of the operation. This systematic review and network meta-analysis (NMA) aimed to systematically assess the efficacy of hemostatic agents in endodontic surgery and to identify the most effective ones. METHODS: PubMed, Scopus, Embase, Cochrane Library, Web of Science, ProQuest, and EBSCOhost databases were searched up to December 2020. We included randomized controlled trials (RCTs) evaluating the efficacy of different hemostatic measures in endodontic surgery, and their risk of bias was assessed using Cochrane's randomized trial tool (RoB 2.0). Frequentist network meta-analysis was conducted, with Odds Ratios and 95% confidence intervals (OR, 95% CI) as effect estimates using the "netmeta" package in R. The quality of evidence was assessed using the CINeMA approach. RESULTS: Six RCTs involving 353 patients (mean age 48.12 y) were included. NMA revealed that aluminum chloride achieved higher hemostatic efficacy than epinephrine (OR = 2.55, 95% CI [1.41, 4.64]), while there was non-significant difference when compared with PTFE strips + epinephrine (OR = 1.00, 95% CI [0.35, 2.90]), electrocauterization (OR = 2.67, 95% CI [0.84, 8.46]), or ferric sulfate (OR = 8.65, 95% CI [0.31, 240.92]). Of all hemostatic agents, aluminum chloride ranked first in control bleeding during endodontic surgery (P-score = 0.84), followed by PTFE strips + epinephrine (P-score = 0.80), electrocauterization (P-score = 0.34), epinephrine (P-score = 0.34), ferric sulfate (P-score = 0.18). The quality of evidence was very low. CONCLUSIONS: Based on the limited data, aluminum chloride provides better hemostasis than epinephrine, while there was no significant difference between the remaining hemostatic agents used in endodontic surgery, which could help clinicians choose the hemostatic agent that achieves adequate hemostasis. achieve adequate hemostasis. Given insufficient evidence, future RCTs addressing this evidence gap are required.

CORONA-Net: Diagnosing COVID-19 from X-ray Images Using Re-Initialization and Classification Networks.

The COVID-19 pandemic has been deemed a global health pandemic. The early detection of COVID-19 is key to combating its outbreak and could help bring this pandemic to an end. One of the biggest challenges in combating COVID-19 is accurate testing for the disease. Utilizing the power of Convolutional Neural Networks (CNNs) to detect COVID-19 from chest X-ray images can help radiologists compare and validate their results with an automated system. In this paper, we propose a carefully designed network, dubbed CORONA-Net, that can accurately detect COVID-19 from chest X-ray images. CORONA-Net is divided into two phases: (1) The reinitialization phase and (2) the classification phase. In the reinitialization phase, the network consists of encoder and decoder networks. The objective of this phase is to train and initialize the encoder and decoder networks by a distribution that comes out of medical images. In the classification phase, the decoder network is removed from CORONA-Net, and the encoder network acts as a backbone network to fine-tune the classification phase based on the learned weights from the reinitialization phase. Extensive experiments were performed on a publicly available dataset, COVIDx, and the results show that CORONA-Net significantly outperforms the current state-of-the-art networks with an overall accuracy of 95.84%.

Safety and Efficacy of Evacetrapib in Patients with Inadequately-controlled Hypercholesterolemia and High Cardiovascular Risk; A meta-analysis of Randomized Placebo-controlled Trials.

BACKGROUND: Low-density lipoprotein cholesterol (LDL-C) is causally related to cardiovascular disease. Inhibition of cholesteryl ester transfer protein with Evacetrapib may provide an additional treatment option for patients who do not reach their LDL-C goal with statins or patients who cannot tolerate statins. We aimed to evaluate the safety and efficacy of Evacetrapib in patients with inadequately-controlled hypercholesterolemia and high cardiovascular risk. METHOD: A computer literature search for PubMed, Scopus, and Science Direct was carried out from inception to 2019 and was updated from January 2019 till March 2021. We included only RCTs. Data were pooled as a mean difference in a random-effect model using the Mantel-Haenzel (M-H) method. We used Open Meta [Analyst] software (by the center of evidence-based medicine, Oxford University, UK). RESULTS: Five studies (n = 12,937 patients) reported in five articles were included in this meta-analysis. The overall pooled estimate showed that LDL-C was significantly lower in the evacetrapib group than the placebo group (MD -34.07 mg/dL, 95% CI [-40.66, -27.49], p<0.0001). The pooled estimate showed that Apo-B was significantly lower in the evacetrapib130 mg group than the placebo group (MD -22.64 mg/dL, 95% CI [-30.70, -14.58], p<0.0001). HDL-C was significantly higher in the evacetrapib group over the placebo group (MD 93.31 mg/dL, 95% CI [56.07, 130.56], p<0.0001). CONCLUSION: Current evidence from five RCTs (12,539 participants) suggests that evacetrapib has favorable outcomes in patients with inadequately-controlled Hypercholesterolemia and high cardiovascular risks. Evacetrapib could significantly increase the HDL and Apo-A1 levels and lower the LDL cholesterol and Apo-B levels with an acceptable safety profile.

Bankart Repair Versus Latarjet Procedure for Recurrent Anterior Shoulder Instability: A Systematic Review and Meta-analysis of 3275 Shoulders.

BACKGROUND: Little consensus is available regarding the standard treatment for recurrent anterior instability of the shoulder. Typically, treatment selection has been based on training and tradition rather than the available evidence. PURPOSE: This study aimed to compare the clinical outcomes between arthroscopic Bankart procedure and the Latarjet procedure in the treatment of recurrent anterior shoulder instability with emphasis on follow-up time. STUDY DESIGN: Systematic review and meta-analysis. METHODS: We searched PubMed, Cochrane Central Register of Controlled Trials, Scopus, Ovid, and Web of Science up to January 2018 and included studies that compared arthroscopic Bankart versus Latarjet for treatment of anterior shoulder instability. Continuous data, such as operative time and patient-reported outcomes, were pooled as mean differences (MDs), whereas dichotomous data, such as recurrence, revision, redislocation, arthropathy, infection, and hematoma, were pooled as risk ratios (RRs), with 95% CIs. RESULTS: Pooling data from 7 cohort studies (3275 patients) showed that arthroscopic Bankart was associated with a higher risk of redislocation (RR, 2.74; 95% CI, 1.48-5.06; P = .03), a higher risk of recurrence (RR, 2.87; 95% CI, 1.91-4.30; P < .0001), and a lower risk of infection (RR, 0.16; 95% CI, 0.06-0.43; P = .0002) compared with Latarjet, while the effect size did not favor arthroscopic Bankart or Latarjet in terms of Rowe score (MD, 0.22; 95% CI, -5.64 to 6.08; P = .94), revision (RR, 0.34; 95% CI, 0.08-1.39; P = .13), and hematoma (RR, 0.20; 95% CI, 0.03-1.19; P = .07). The effect estimate showed a pronounced advantage for Latarjet from 6 to 10 years postoperatively in terms of recurrence and redislocation (RR, 3.00; 95% CI, 1.98-4.56 and RR, 2.85; 95% CI, 1.51-5.38, respectively). CONCLUSION: Our results showed that Latarjet had less risk of recurrence and redislocation with longer follow-up time. Both procedures were comparable in terms of Rowe score, the need for revision, and postoperative hematoma formation, whereas Bankart repair was associated with a lower risk of infection.

Deep Learning-Based Crowd Scene Analysis Survey.

Recently, our world witnessed major events that attracted a lot of attention towards the importance of automatic crowd scene analysis. For example, the COVID-19 breakout and public events require an automatic system to manage, count, secure, and track a crowd that shares the same area. However, analyzing crowd scenes is very challenging due to heavy occlusion, complex behaviors, and posture changes. This paper surveys deep learning-based methods for analyzing crowded scenes. The reviewed methods are categorized as (1) crowd counting and (2) crowd actions recognition. Moreover, crowd scene datasets are surveyed. In additional to the above surveys, this paper proposes an evaluation metric for crowd scene analysis methods. This metric estimates the difference between calculated crowed count and actual count in crowd scene videos.

A novel fully convolutional network for visual saliency prediction.

A human Visual System (HVS) has the ability to pay visual attention, which is one of the many functions of the HVS. Despite the many advancements being made in visual saliency prediction, there continues to be room for improvement. Deep learning has recently been used to deal with this task. This study proposes a novel deep learning model based on a Fully Convolutional Network (FCN) architecture. The proposed model is trained in an end-to-end style and designed to predict visual saliency. The entire proposed model is fully training style from scratch to extract distinguishing features. The proposed model is evaluated using several benchmark datasets, such as MIT300, MIT1003, TORONTO, and DUT-OMRON. The quantitative and qualitative experiment analyses demonstrate that the proposed model achieves superior performance for predicting visual saliency.

DP-Siam: Dynamic Policy Siamese Network for Robust Object Tracking.

Balancing the trade-off between real-time performance and accuracy in object tracking is a major challenge. In this paper, a novel dynamic policy gradient Agent-Environment architecture with Siamese network (DP-Siam) is proposed to train the tracker to increase the accuracy and the expected average overlap while performing in real-time. DP-Siam is trained offline with reinforcement learning to produce a continuous action that predicts the optimal object location. DP-Siam has a novel architecture that consists of three networks: an Agent network to predict the optimal state (bounding box) of the object being tracked, an Environment network to get the Q-value during the offline training phase to minimize the error of the loss function, and a Siamese network to produce a heat-map. During online tracking, the Environment network acts as a verifier to the Agent network action. Extensive experiments are performed on six widely used benchmarks: OTB2013, OTB50, OTB100, VOT2015, VOT2016 and VOT2018. The results show that DP-Siam significantly outperforms the current state-of-the-art trackers.

Historically, did Cemented Thompson perform better than uncemented Austin Moore hemiarthroplasty for femoral neck fractures? A meta-analysis of available evidence.

INTRODUCTION: Thompson and Austin Moore prostheses have been commonly used in hemiarthroplasties for displaced femoral neck fractures. There has been considerable debate about which of these prostheses is preferred. The purpose of this meta-analysis was to compare historical data for clinical outcomes of cemented Thompson and uncemented Austin Moore hemiarthroplasty in displaced femoral neck fractures. METHODS: We searched Medline via PubMed, Cochrane Central, Scopus, Ovid and Web of Science for relevant articles up to February 2019. The included outcomes measured were hip function, hip pain, implant-related complications, surgical complications, reoperation rate and hospital stay. The data were pooled as risk ratio (RR) or mean difference (MD) with 95% confidence interval (CI) between the two compared groups in a meta-analysis model. RESULTS: Ten studies (four RCTs and six observational studies) with a total of 4378 patients were included in the final analysis. The pooled RR showed that the Thompson group was associated with a lower incidence of postoperative hip pain (RR = 0.66, 95% CI [0.54, 0.80]), lesser reoperation rate (RR = 0.46, 95% CI [0.24, 0.88]), lesser intraoperative fractures (RR = 0.15, 95% CI [0.09, 0.25]), but a longer operative time (MD = 12.04 min, 95% CI [2.08, 22.00]) in comparison to the Austin Moore group. The effect estimate did not favour either group in terms of hip function, periprosthetic fractures, prosthetic dislocations, wound infection, mortality and hospital stay. CONCLUSION: Evidence shows that Thompson hemiarthroplasty is better than Austin Moore hemiarthroplasty in terms of hip pain, reoperation rate and intraoperative fractures. Whereas the postoperative hip function is equivalent, these results could be considered when assessing the outcomes in modern hips.

Sliding hip screws versus cancellous screws for femoral neck fractures: a systematic review and meta-analysis.

PURPOSE: Both sliding hip screws (SHS) and cancellous screws are used in the surgical management of intracapsular femoral neck fracture. However, there is paucity of information as to which is the superior treatment modality. We performed this systematic review and meta-analysis study to compare the clinical outcomes of SHS and cancellous screws for the treatment of femoral neck fractures in adult patients. METHODS: We searched PubMed, Scopus, Web of Science, and Cochrane CENTRAL, up to December 2017. Randomized controlled trials (RCTs) directly comparing the clinical outcomes of SHS and cancellous screws for femoral neck fractures were retrieved with no language or publication year restrictions. Data retrieved included operative details, nonunion rate, avascular necrosis, reoperation, infection and mortality, hip pain, functional hip scores, and medical complications. These were pooled as risk ratio or mean difference (MD) with their corresponding 95% confidence interval (CI). Heterogeneity was assessed by Chi-square test. RESULTS: Ten RCTs involving 1934 patients were included in the final analysis. The pooled estimate showed that the SHS group was associated with more intraoperative blood loss (MD = 110.01 ml, 95% CI [52.42, 167.60], p = 0.00002) than the cancellous screws. There was no significant difference in terms of operative time, postoperative hip function, nonunion, avascular necrosis, reoperation rate, infection, fracture healing, hip pain, medical complications, and mortality rate. CONCLUSION: Based on our study, the cancellous screws group was associated with less intraoperative blood loss in comparison with the SHS group. No other significant differences were found between the two interventions.

An Accelerated Sequential PCP-Based Method for Ground-Moving Objects Detection From Aerial Videos.

The detection of ground-moving objects in aerial videos has evolved over the years to handle more challenges such as large camera motion, the small size of the objects, and occlusion. Recently, aerial detection has been attempted using principal component pursuit (PCP) due to its superiority in detecting small moving objects. However, PCP-based detection methods generally suffer from high-false detections as well as high-computational loads. This paper presents a novel PCP-based detection method called kinematic regularization with local null space pursuit (KRLNSP) that drastically reduces false detections and the computational loads. KRLNSP models the background in an aerial video as a subspace that spans a low-dimension subspace while it models the moving objects as moving sparse. Accordingly, the detection is achieved by using multiple local null spaces and enhanced kinematic regularization. The multiple local null spaces allow real-time execution to nullify the background while preserving the moving objects unchanged. The kinematic regularization penalizes these moving objects to filter out false detections. The extensive evaluation of KRLNSP and relevant current state-of-the-art methods prove that the KRLNSP outperforms these methods (the true positive rate of KRLNSP is 98% and its false positive rate is 0.4%) and significantly reduces the computational loads (KRLNSP execution time is 0.3 s/frame).

Correction to: Contemporary cemented versus uncemented hemiarthroplasty for the treatment of displaced intracapsular hip fractures: a meta-analysis of forty-two thousand forty-six hips.

The author name "Muhammad Shawqi" was incorrectly presented as "Muhammad Shawq".

Contemporary cemented versus uncemented hemiarthroplasty for the treatment of displaced intracapsular hip fractures: a meta-analysis of forty-two thousand forty-six hips.

INTRODUCTION: Controversy exists regarding the use of cement for hemiarthroplasty to treat displaced intracapsular hip fractures. The aim of this systematic review and meta-analysis was to compare the clinical outcomes between contemporary cemented and contemporary uncemented hemiarthroplasty for the treatment of displaced femoral neck fractures. METHODS: Literature searches of PubMed, Scopus, Web of Science, and Cochrane Central, up to May 2017, were performed. We included randomized controlled trials (RCTs) and observational studies comparing contemporary cemented with contemporary uncemented hemiarthroplasty. Data were pooled as mean difference (MD) or risk ratio (RR) with 95% confidence interval (CI) in a meta-analysis model. Studies with the Thompson and Austin Moore prostheses were excluded. RESULTS: A total of 29 studies (9 RCTs and 20 observational studies), with a total of 42,046 hips, were included. Meta-analysis showed that the cemented group was associated with fewer periprosthetic fractures (RR = 0.44, 95% CI [0.21, 0.91]), longer operative time (MD = 11.25 min, 95% CI [9.85, 12.66]), more intraoperative blood loss (MD = 68.72 ml, 95% CI [50.76, 86.69]), and higher heterotopic ossification (RR = 1.79, 95% CI [1.11, 2.88]) compared with the uncemented group. Meta-analysis showed no significant difference in terms of post-operative hip function, hip pain, reoperation rate, prosthetic dislocations, aseptic loosening, wound infection, and hospital stay. CONCLUSIONS: This meta-analysis shows that contemporary cemented prostheses have less intra-operative and post-operative fractures, but longer operative time, more intra-operative blood loss, and heterotopic ossifications. Otherwise, there were no significant differences between both groups.

Efficacy and safety of tramadol in pain relief during diagnostic outpatient hysteroscopy: systematic review and meta-analysis of randomized controlled trials.

OBJECTIVE: To evaluate the evidence about the safety and efficacy of tramadol in pain relief during diagnostic outpatient hysteroscopy (OH). DESIGN: Systematic review and meta-analysis of randomized controlled trials. SETTING: Not applicable. PATIENT(S): Patients undergoing diagnostic OH received tramadol versus those who were administered placebo. INTERVENTION(S): Electronic databases were searched using the following MeSH terms (tramadol OR opioids OR opioid OR narcotic OR narcotics) AND (hysteroscopy OR hysteroscopic). MAIN OUTCOME MEASURE(S): Pain assessed by visual analogue scale (VAS) during OH, immediately and 30 minutes after the procedure. RESULT(S): One hundred thirteen studies were identified of which four randomized clinical trials were deemed eligible for this review (tramadol: n = 209; placebo: n = 209). The pooled estimate showed that tramadol significantly reduced VAS during procedure than placebo (weighted mean difference [WMD] = -1.33; 95% confidence interval [CI] -1.78 to -0.88, I2 = 3%, P = .36). In addition, tramadol significantly reduced VAS immediately after the procedure (WMD = -1.05; 95% CI -1.49 to -0.61, I2= 0, P = .84) and 30 minutes after (WMD = -0.98; 95% CI -1.30 to -0.65, I2 = 0, P = .88). CONCLUSION(S): This meta-analysis suggests that tramadol is safe, effective, and gives favorable results in reducing pain during diagnostic OH.

Semi-Automatic Algorithms for Estimation and Tracking of AP-Diameter of the IVC in Ultrasound Images.

Acutely ill patients presenting with conditions such as sepsis, trauma, and congestive heart failure require judicious resuscitation in order to achieve and maintain optimal circulating blood volume. Increasingly, emergency and critical care physicians are using portable ultrasound to approximate the temporal changes of the anterior-posterior (AP)-diameter of the inferior vena cava (IVC) in order to guide fluid administration or removal. This paper proposes semi-automatic active ellipse and rectangle algorithms capable of improved and quantified measurement of the AP-diameter. The proposed algorithms are compared to manual measurement and a previously published active circle model. Results demonstrate that the rectangle model outperforms both active circle and ellipse irrespective of IVC shape and closely approximates tedious expert assessment.

Estimation and tracking of AP-diameter of the inferior vena cava in ultrasound images using a novel active circle algorithm.

Medical research suggests that the anterior-posterior (AP)-diameter of the inferior vena cava (IVC) and its associated temporal variation as imaged by bedside ultrasound is useful in guiding fluid resuscitation of the critically-ill patient. Unfortunately, indistinct edges and gaps in vessel walls are frequently present which impede accurate estimation of the IVC AP-diameter for both human operators and segmentation algorithms. The majority of research involving use of the IVC to guide fluid resuscitation involves manual measurement of the maximum and minimum AP-diameter as it varies over time. This effort proposes using a time-varying circle fitted inside the typically ellipsoid IVC as an efficient, consistent and novel approach to tracking and approximating the AP-diameter even in the context of poor image quality. In this active-circle algorithm, a novel evolution functional is proposed and shown to be a useful tool for ultrasound image processing. The proposed algorithm is compared with an expert manual measurement, and state-of-the-art relevant algorithms. It is shown that the algorithm outperforms other techniques and performs very close to manual measurement.