Full Video Research Articles

AMPS: Predicting popularity of short-form videos using multi-modal attention mechanisms in social media marketing environments

Emerging as a dominant content format amid the shift from television to mobile, short-form videos wield immense potential across diverse domains. However, the scarcity of datasets and established metrics for their popularity evaluation poses a challenge in accurately reflecting their real-world distribution. In response, our work introduces a dataset and pioneers a cumulative distribution function-based standard tailored specifically for short-form videos. Our model, AMPS (Attention-based Multi-modal Popularity prediction model of Short-form videos) is designed to effectively forecast the popularity of these videos. Considering YouTube Shorts, typically confined to under one minute, our research capitalizes on complete video frames for a holistic prediction of popularity. AMPS harnesses BiLSTM with Self-Attention and Co-Attention mechanisms, enabling a deeper understanding of intra-modal and inter-modal relationships across various modalities. Leveraging full video frame representation, our model significantly enhances prediction accuracy. Comprehensive evaluations against baseline models and machine learning algorithms consistently showcase AMPS' superiority in metrics like G-Mean, average F1-score, and Accuracy. Furthermore, when compared with other open social media datasets, our dataset coupled with AMPS consistently outperforms, affirming its robustness and reliability. Additionally, ablation studies underscore the effectiveness of AMPS' architecture and highlight the significance of each modality in predicting popularity.

Analysis of neural networks for routine classification of sixteen ultrasound upper abdominal cross sections.

Abdominal ultrasound screening requires the capture of multiple standardized plane views as per clinical guidelines. Currently, the extent of adherence to such guidelines is dependent entirely on the skills of the sonographer. The use of neural network classification has the potential to better standardize captured plane views and streamline plane capture reducing the time burden on operators by combatting operator variability. A dataset consisting of 16 routine upper abdominal ultrasound scans from 64 patients was used to test the classification accuracy of 9 neural networks. These networks were tested on both a small, idealised subset of 800 samples as well as full video sweeps of the region of interest using stratified sampling and transfer learning. The highest validation accuracy attained by both GoogLeNet and InceptionV3 is 83.9% using transfer learning and the large sample set of 26,294 images. A top-2 accuracy of 95.1% was achieved using InceptionV3. Alexnet attained the highest accuracy of 79.5% (top-2 of 91.5%) for the smaller sample set of 800 images. The neural networks evaluated during this study were also successfully able to identify problematic individual cross sections such as between kidneys, with right and left kidney being accurately identified 78.6% and 89.7%, respectively. Dataset size proved a more important factor in determining accuracy than network selection with more complex neural networks providing higher accuracy as dataset size increases and simpler linear neural networks providing better results where the dataset is small.

Computer-Based Listening Test with Full Video, Visual-Limited Video, and Audio: A Comparative Analysis Based on Difficulty, Discrimination Power, and Response Time

ABSTRACT This study aimed to compare item characteristics and response time between stimulus conditions in computer-delivered listening tests. Listening materials had three variants: regular videos, frame-by-frame videos, and only audios without visuals. Participants were 228 Japanese high school students who were requested to complete one of nine listening test forms. Results showed that item difficulty did not significantly differ between conditions, but item discrimination was much weaker in regular video versions than in the other two conditions. Response time was not different in the three conditions in any quantile. This suggests that listening items full of visual information have different aspects of the general (traditional) listening construct, despite the possibility of construct-irrelevant factors being affected.

Rich Action-semantic Consistent Knowledge for Early Action Prediction.

Early action prediction (EAP) aims to recognize human actions from a part of action execution in ongoing videos, which is an important task for many practical applications. Most prior works treat partial or full videos as a whole, ignoring rich action knowledge hidden in videos, i.e., semantic consistencies among different partial videos. In contrast, we partition original partial or full videos to form a new series of partial videos and mine the Action-Semantic Consistent Knowledge (ASCK) among these new partial videos evolving in arbitrary progress levels. Moreover, a novel Rich Action-semantic Consistent Knowledge network (RACK) under the teacher-student framework is proposed for EAP. Firstly, we use a two-stream pre-trained model to extract features of videos. Secondly, we treat the RGB or flow features of the partial videos as nodes and their action semantic consistencies as edges. Next, we build a bi-directional semantic graph for the teacher network and a single-directional semantic graph for the student network to model rich ASCK among partial videos. The MSE and MMD losses are incorporated as our distillation loss to enrich the ASCK of partial videos from the teacher to the student network. Finally, we obtain the final prediction by summering the logits of different subnetworks and applying a softmax layer. Extensive experiments and ablative studies have been conducted, demonstrating the effectiveness of modeling rich ASCK for EAP. With the proposed RACK, we have achieved state-of-the-art performance on three benchmarks. The code is available at https://github.com/lily2lab/RACK.git.

MAR: Masked Autoencoders for Efficient Action Recognition

Standard approaches for video action recognition usually operate on full input videos, which is inefficient due to the widespread spatio-temporal redundancy in videos. The recent progress in masked video modelling, specifically VideoMAE, has shown the ability of vanilla Vision Transformers (ViT) to complement spatio-temporal contexts using limited visual content. Inspired by this, we propose Masked Action Recognition (MAR), which reduces redundant computation by discarding a proportion of patches and operating only on a portion of the videos. MAR includes two essential components: <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">cell running masking</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">bridging classifier</i> . Specifically, to enable the ViT to perceive the details beyond the visible patches, cell running masking is used to preserve the spatio-temporal correlations in videos. This ensures that the patches at the same spatial location can be observed in turn for easy reconstructions. Additionally, we notice that, although the partially observed features can reconstruct semantically explicit invisible patches, they fail to achieve accurate classification. To address this issue, we propose a bridging classifier that can help fill the semantic gap between the ViT encoded features used for reconstruction and the specialized features used for classification. Our proposed MAR can reduce the computational cost of ViT by 53%. Extensive experiments have demonstrated that MAR consistently outperforms existing ViT models by a notable margin. Notably, we found that a ViT-Large model fine-tuned by MAR achieves comparable performance to a ViT-Huge model fine-tuned by standard training methods on both Kinetics-400 and Something-Something v2 datasets. Moreover, the computation overhead of our ViT-Large model is only 14.5% of that of the ViT-Huge model. Codes have been made available <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/alibaba-mmai-research/Masked-Action-Recognition</uri> .

Sonographically-navigated frozen shoulder release (S-FSR): A modified technique for sonographically-navigated, in-office hydrodilatation of adhesive capsulitis

We describe a new technique for hydrodilatation of the frozen shoulder, which we coined ‘Sonographically-Navigated Frozen Shoulder Release (S-FSR)’ or ‘Dr Gonzalez's technique.’ Traditional treatments include a combination of conservative and surgical modalities, such as non-steroidal anti-inflammatories, physical therapy, and open capsular release. We describe a modification to hydrodilatation of the frozen shoulder. Our technique describes the gradual dilation of the glenohumeral (GH) capsule with the goal of seeing a gentle release of the frozen shoulder. Furthermore, we outline our protocol for patient preparation with preprocedural diazepam 5 mg and Hydrocodone 5 mg-Acetaminophen 325 mg, one tablet each. During the procedure, we inject a solution of 10 mL lidocaine mixed with 2 mL of 40 mg/mL triamcinolone acetonide injection (80 mg total) through an anterior approach at the level of the rotator interval for pain control. Following this injection, we proceed to fenestrate the superior glenohumeral ligament (SGHL) in the process. We conclude the procedure with hydrodilatation of the GH joint (GHJ) through the posterior approach with approximately 50 mL of 0.9% normal saline, or until an expansion and release of the GH joint is visualized under ultrasound visualization. A full instruction video can be found at: https://www.youtube.com/watch?v = ZNB0R0hkeok&amp;ab_channel = DrJorgeA.Gonzalez

3D far-field Lidar sensing and computational modeling for human identification.

3D sensors offer depth sensing that may be used for task-specific data processing and computational modeling. Many existing methods for human identification using 3D depth sensors primarily focus on Kinect data, where the range is very limited. This work considers a 3D long-range Lidar sensor for far-field imaging of human subjects in 3D Lidar full motion video (FMV) of "walking" action. 3D Lidar FMV data for human subjects are used to develop computational modeling for automated human silhouette and skeleton extraction followed by subject identification. We propose a matrix completion algorithm to handle missing data in 3D FMV due to self-occlusion and occlusion from other subjects for 3D skeleton extraction. We further study the effect of noise in the 3D low resolution far-field Lidar data in human silhouette extraction performance of the model. Moreover, this work addresses challenges associated with far-field 3D Lidar including learning with a limited amount of data and low resolution. Moreover, we evaluate the proposed computational algorithm using a gallery of 10 subjects for human identification and show that our method is competitive with the state-of-the-art OpenPose and V2VPose skeleton extraction models using the same dataset for human identification.

Abstract 11685: Effects of Pectus Excavatum Surgical Repair on Cardiac Volumes and Systolic Function as Assessed by 3D Transesophageal Echocardiography

Introduction: The consequences of external cardiovascular compression and cardiac displacement from pectus excavatum (PEx) are debated. The complex geometry of the right (RV) and left ventricle (LV) along with distortion from the extrinsic chest deformity pose significant limitations to reliable assessment of LV and RV volumes and function using transthoracic 2D-echocardiography in this population. Hypothesis: A 3D-echocardiography approach can reliably assess the effects of external cardiac compression and the benefits of surgical correction on RV and LV function in PEx patients. Methods: Adult patients undergoing minimally invasive repair of PEx between Oct-2021 and Feb-2023 were included in this prospective study. Intraoperative transesophageal (TEE) 6 beat full volume video loops incorporating LV and RV were captured and compared before and after surgical correction of the PEx. TomTec Software® was used for offline 3D/4D reconstruction ( Figure 1 shows changes in 3D RV and LV shape pre- and post-surgery). Result: The study included 102 patients (mean age of 32.1 ±11.8 years; Haller index 4.9 ±2.5; 60.8% males). Post PEx repair, there was a significant (p&lt;0.001) increase in the 3D RV volumes, RV ejection fraction (EF), RV stroke volume and RV strain ( Figure 1 ). Additionally, there was a significant improvement in 3D LVEF, 3D LV strain (longitudinal and circumferential), LV twist, and torsion ( Figure 1 ). Conclusion: In this PEx cohort, 3D/4D TEE imaging demonstrated significant increases in RV volumes and in RV and LV systolic function after surgical repair. These findings support the negative cardiovascular consequences of PEx compression and benefits of surgical repair.

An Analysis of Positive Politeness Strategies on You Tube Video Entitled “Ariana Grande I Full Video”

The aim of this research is to find and explain the occurrence of positive politeness strategies from You Tube video entitled “Ariana Grande I Full Video”. In analyzing the data, researchers took the theory of positive politeness strategies in the scope of Pragmatics by Brown &amp; Levinson (1987) that consists 15 strategies. Descriptive qualitative method was applied by researchers in this research. In collecting data, researchers took observational method, such as watching and listening careful to the video, finding the words that contain the occurrence of positive politeness, collecting the data from the video, divide them based on positive politeness strategies, analyze the words, and explain the result of the analysis. After done doing the analysis, researchers found 68 phenomenon that occur in 10 positive politeness strategies, they are 3 occurrence of the first strategy (Notice, attending to hearer), 13 occurrence of the second strategy (Exaggerating (Showing interest, approving, and give sympathy to hearer)), 8 occurrence of the fourth strategy (Using in group identity marker), 21 occurrence of the fifth strategy (Seeking agreement), 3 occurrence of the sixth strategy (Avoid disagreement), 10 occurrence of the seventh strategy (Presupposing, raising, or asserting common ground), 1 occurrence of the eighth strategy (Giving joke), 1 occurrence of the twelfth strategy (Including or getting both speaker and hearer into the activity), 7 occurrence of the thirteenth strategy (Giving or asking for reason), and 1 occurrence of the fifteenth strategy (Giving gifts to H in the form of goods, sympathy, understanding, and cooperation).

Video Summarization Based on Feature Fusion and Data Augmentation

During the last few years, several technological advances have led to an increase in the creation and consumption of audiovisual multimedia content. Users are overexposed to videos via several social media or video sharing websites and mobile phone applications. For efficient browsing, searching, and navigation across several multimedia collections and repositories, e.g., for finding videos that are relevant to a particular topic or interest, this ever-increasing content should be efficiently described by informative yet concise content representations. A common solution to this problem is the construction of a brief summary of a video, which could be presented to the user, instead of the full video, so that she/he could then decide whether to watch or ignore the whole video. Such summaries are ideally more expressive than other alternatives, such as brief textual descriptions or keywords. In this work, the video summarization problem is approached as a supervised classification task, which relies on feature fusion of audio and visual data. Specifically, the goal of this work is to generate dynamic video summaries, i.e., compositions of parts of the original video, which include its most essential video segments, while preserving the original temporal sequence. This work relies on annotated datasets on a per-frame basis, wherein parts of videos are annotated as being “informative” or “noninformative”, with the latter being excluded from the produced summary. The novelties of the proposed approach are, (a) prior to classification, a transfer learning strategy to use deep features from pretrained models is employed. These models have been used as input to the classifiers, making them more intuitive and robust to objectiveness, and (b) the training dataset was augmented by using other publicly available datasets. The proposed approach is evaluated using three datasets of user-generated videos, and it is demonstrated that deep features and data augmentation are able to improve the accuracy of video summaries based on human annotations. Moreover, it is domain independent, could be used on any video, and could be extended to rely on richer feature representations or include other data modalities.

Video-Restoration-Net: Deep Generative Model with Non-Local Network for Inpainting and Super-Resolution Tasks

Although deep learning-based approaches for video processing have been extensively investigated, the lack of generality in network construction makes it challenging for practical applications, particularly in video restoration. As a result, this paper presents a universal video restoration model that can simultaneously tackle video inpainting and super-resolution tasks. The network, called Video-Restoration-Net (VRN), consists of four components: (1) an encoder to extract features from each frame, (2) a non-local network that recombines features from adjacent frames or different locations of a given frame, (3) a decoder to restore the coarse video from the output of a non-local block, and (4) a refinement network to refine the coarse video on the frame level. The framework is trained in a three-step pipeline to improve training stability for both tasks. Specifically, we first suggest an automated technique to generate full video datasets for super-resolution reconstruction and another complete-incomplete video dataset for inpainting, respectively. A VRN is then trained to inpaint the incomplete videos. Meanwhile, the full video datasets are adopted to train another VRN frame-wisely and validate it against authoritative datasets. We show quantitative comparisons with several baseline models, achieving 40.5042 dB/0.99473 on PSNR/SSIM in the inpainting task, while during the SR task we obtained 28.41 dB/0.7953 and 27.25/0.8152 on BSD100 and Urban100, respectively. The qualitative comparisons demonstrate that our proposed model is able to complete masked regions and implement super-resolution reconstruction in videos of high quality. Furthermore, the above results show that our method has greater versatility both in video inpainting and super-resolution tasks compared to recent models.

Introduction: 2022 Daniel H. Wagner Prize for Excellence in the Practice of Advanced Analytics and Operations Research

The judges for the 2022 Daniel H. Wagner Prize for Excellence in the Practice of Advanced Analytics and Operations Research selected the four finalist papers featured in this special issue of the INFORMS Journal on Applied Analytics (IJAA). The prestigious Wagner Prize—awarded for achievement in implemented operations research, management science, and advanced analytics—emphasizes the quality and originality of mathematical models along with clarity of written and oral exposition. This year’s winning application describes the design and deployment of a generalized synthetic control, a powerful and innovative statistical method for identifying, in a noisy environment, retailing innovations that produce a small percentage improvement in a large volume of sales for Anheuser Busch Inbev. The remaining three papers describe an inverse control approach to allocating lung transplants that best meets targeted outcomes and has been implemented as the national lung allocation policy on March 9, 2023, across the United States; a human-centric, optimized parcel delivery system developed for Deutsche Post that saves money while meeting constraints learned dynamically from driver behavior; and an AI-based system developed for Alibaba that learns supplier behavior to improve replenishment ordering and inventory control. Supplemental Material: Full presentation videos with slides are available in the INFORMS Video Library at https://www.informs.org/Resource-Center/Video-Library and as electronic companions to the INFORMS Journal on Applied Analytics articles.

RESEARCH ON WIDE SPECTRUM FULL-COLOR NIGHT VISION ELDERLY FACIAL VIDEO HEART RATE AUTOMATIC MEASUREMENT AND WARNING SYSTEM

In recent years, the prevalence of cardiovascular diseases has been growing geometrically, especially among the elderly. Real-time heart rate monitoring becomes an effective measure for preventing cardiovascular diseases. Traditional contact-based real-time heart rate measurement methods require long-term wearing of contact electrodes, which can cause inconvenience to the elderly population. In this study, a wide spectrum full color night vision elderly face video heart rate automatic measurement and early warning system was designed and developed by combining imaging photoplethysmography (IPPG), video acquisition equipment based on starlight wide spectrum camera and high-speed face recognition signal processing algorithm based on deep learning. By constructing a software and hardware platform, a comprehensive evaluation of the system’s application in elderly facial video heart rate measurement experiments was conducted.

Evaluation of online videos to engage viewers and support decision-making for COVID-19 vaccination: how narratives and race/ethnicity enhance viewer experiences.

Vaccine hesitancy has hampered the control of COVID-19 and other vaccine-preventable diseases. We conducted a national internet-based, quasi-experimental study to evaluate COVID-19 vaccine informational videos. Participants received an informational animated video paired with the randomized assignment of (1) a credible source (differing race/ethnicity) and (2) sequencing of a personal narrative before or after the video addressing their primary vaccine concern. We examined viewing time and asked video evaluation questions to those who viewed the full video. Among 14,235 participants, 2,422 (17.0%) viewed the full video. Those who viewed a personal story first (concern video second) were 10 times more likely to view the full video (p < 0.01). Respondent-provider race/ethnicity congruence was associated with increased odds of viewing the full video (aOR: 1.89, p < 0.01). Most viewers rated the informational video(s) to be helpful, easy to understand, trustworthy, and likely to impact others' vaccine decisions, with differences by demographics and also vaccine intentions and concerns. Using peer-delivered, personal narrative, and/or racially congruent credible sources to introduce and deliver vaccine safety information may improve the openness of vaccine message recipients to messages and engagement.

A Closer Look on Indonesian EFL students’ Writing Process: The Application of Cognitive and Metacognitive Strategies

Cognitive and metacognitive strategies have been researched to find an effective strategy in teaching writing. However, little attention has been given to investigate what is on the students’ minds while applying those strategies in the writing process. This descriptive qualitative study aims to investigate the process of applying the cognitive and metacognitive strategies by EFL students while performing argumentative essays. Three university students of the English Department taking the Essay Writing subject were invited to record the process of authoring argumentative essays using Zoom recording to display full audio and video on screen. Their writings were checked and the students were interviewed. This process used think-aloud protocols (TAPs) completed with the screen recording method (SRM) to record all activities on the computer screen, students’ work, and the interview explored the learners’ perception towards particular situation in their writing process. The data were categorized and analyzed using English as a foreign language (EFL) writing strategies as a framework for narrative analysis. This study found that both cognitive and metacognitive strategies helped all participants complete their essay through the thinking process and decision making of each step and strategy. While cognitive and metacognitive were applied distinctively in each participant’s writing processes, it produced the internalization of writing steps critically in self-regulated learning. This study demonstrates that cognitive and metacognitive processes are effective in promoting varied writing strategies and self-regulated learning for EFL learners. To get a thorough application of cognitive and metacognitive strategies in writing activity, future research can use other genre for students at different age or level of education.

Focal plane array based on HgTe nanocrystals with photovoltaic operation in the short-wave infrared

Thanks to their remarkable spectral tunability across the entire infrared range, HgTe nanocrystals present a unique platform for designing infrared optoelectronic devices. While in recent years most of the significant advances in this domain have been made on devices at the single-pixel level, there is a growing trend toward exploring the potential of this material for imaging applications. However, until recently, focal plane arrays based on HgTe colloidal nanocrystals have been limited to the photoconductive mode, which is inherently associated with a large dark current. In this work, we demonstrate a diode stack compatible with a readout integrated circuit whose back-end processing has been optimized to ensure compatibility with a complete diode stack deposition. The diode design is also optimized to generate a Fabry–Pérot cavity in which 50% of the light is effectively absorbed at the band edge. Finally, taking benefit from the full video graphics array format, high-resolution images are taken.

O-158 Prediction of implantation of vitrified-warmed blastocysts using a deep learning algorithm on a single post-warming image of each embryo

Abstract Study question Can a deep learning algorithm determine the implantation likelihood of vitrified blastocysts from single time-point post-warming time-lapse images obtained from EmbryoScope? Summary answer A significant positive correlation was found between the score given by the deep learning algorithm and the implantation rate of the frozen blastocysts analyzed. What is known already Despite the safety and success of the vitrification technique, many of the best morphologically classified embryos prior to vitrification do not maintain their characteristics when returned to room temperature. Since artificial intelligence (AI) has been shown to be useful in assessing the quality of fresh embryos at different stages, it could also be used to predict the outcome of frozen embryo transfers. This is the first attempt to distinguish vitrified blastocysts that will implant from those that will not by applying deep learning on a single image taken in the period between warming and transfer. Study design, size, duration This retrospective single-centre study included 689 blastocysts with known implantation data. All of them were vitrified and warmed using the Cryotop method (Kitazato, Biopharma, Japan). Warmed embryos were assessed by experienced embryologists according to the ASEBIR morphological scoring system and the degree of expansion. Immediately after warming, blastocysts were placed in EmbryoScope (Vitrolife, Denmark) time-lapse incubators for 2-5 hours to let them re-expand until embryo transfer. Participants/materials, setting, methods A pre-transfer embryo image was provided to the algorithm (Life Whisperer Viability), which returned scores from 0 to 10. Data were analyzed by chi-square test or one-way ANOVA. Multivariate logistic regression analysis was performed including embryo score, oocyte origin (donated vs. autologous), oocyte age, patient age, oocyte handling (fresh vs. vitrified eggs), and day of vitrification (5 vs. 6). The area under the receiver operating characteristic (ROC) curve (AUC) was used to calculate evaluation performance. Main results and the role of chance There was a significant difference between the scores given to vitrified implanted and non-implanted embryos. The mean value was 4.98 [95% CI 4.70-5.26]* for non-implanted embryos and 6.16 [95% CI 5.84-6.48]* for implanted embryos. When dividing the embryo score by quartiles, the implantation rate for each quartile was 27.7% for Q1, 33.7% for Q2 (Odds ratio (OR)=1.31 [0.82–2.09]), 44.8% for Q3 (OR = 2.05 [1.30–3.22]*) and 48.3% for Q4 (OR = 2.46 [1.56–3.87]*). The contribution of each increased unit of the algorithm score to the implantation (OR = 1.16 [95% CI: 1.09–1.23]*) was statistically significant. The deep learning algorithm score successfully predicted implantation with an AUC of 0.65 [95% CI: 0.60–0.69]*. Moreover, the algorithm performed better than the ASEBIR morphological classification, which achieved an AUC of 0,62 [95% CI: 0.58–0.66] in predicting implantation. *P &amp;lt; 0.001. Limitations, reasons for caution Our clinic was not involved in the development of the algorithm. Furthermore, it was designed to assess fresh embryos prior to vitrification but not thawed ones, so this study should be considered an external trial. A post-warming full video analysis by AI could possibly provide additional information. Wider implications of the findings The use of predictive models on vitrified cycles may help to select the best frozen embryo for transfer and predict which embryos will result in implantation failure early enough to thaw another one with a better chance of success. It also indicates that AI can generalize to thawed embryo assessment. Trial registration number -

Video Surveillance System and Violence Detection

The Video Surveillance System is an enhanced version of modern CCTV surveillance. The system is trained using Yolo algorithm to detect the weapons and also is trained using Convolutional Neural Network (CNN) to detect Violence in public places. The system not only detects but also it sends notification to the central control of the system like Police Stations and the police will be able to take necessary actions. The users can also send complaints and use the android application in case of emergency to alert the police if the user is in danger situations. The police will be able to take quick actions using the locations available with the complaint or emergency. By combining CCTV surveillance and weapon detection the system can act as a Central Surveillance System. The system will not record the full video instead take the snapshot of the detected weapons or violence and send it to the admin saving the memory.

Two- Versus 8-Zone Lung Ultrasound in Heart Failure: Analysis of a Large Data Set Using a Deep Learning Algorithm.

Scanning protocols for lung ultrasound often include 8 or more lung zones, which may limit real-world clinical use. We sought to compare a 2-zone, anterior-superior thoracic ultrasound protocol for B-line artifact detection with an 8-zone approach in patients with known or suspected heart failure using a deep learning (DL) algorithm. Adult patients with suspected heart failure and B-lines on initial lung ultrasound were enrolled in a prospective observational study. Subjects received daily ultrasounds with a hand-held ultrasound system using an 8-zone protocol (right and left anterior/lateral and superior/inferior). A previously published deep learning algorithm that rates severity of B-lines on a 0-4 scale was adapted for use on hand-held ultrasound full video loops. Average severities for 8 and 2 zones were calculated utilizing DL ratings. Bland-Altman plot analyses were used to assess agreement and identify bias between 2- and 8-zone scores for both primary (all patients, 5728 videos, 205 subjects) and subgroup (confirmed diagnosis of heart failure or pulmonary edema, 4464 videos, 147 subjects) analyses. Bland-Altman plot analyses revealed excellent agreement for both primary and subgroup analyses. The absolute difference on the 4-point scale between 8- and 2-zone average scores was not significant for the primary dataset (0.03; 95% CI -0.01 to 0.07) or the subgroup (0.01; 95% CI -0.04 to 0.06). Utilization of a 2-zone, anterior-superior thoracic ultrasound protocol provided similar severity information to an 8-zone approach for a dataset of subjects with known or suspected heart failure.

A teaching video for abdominal myomectomy with Maylard incision

Myomectomy is a common surgical procedure performed by Obstetricians and Gynecologists on women during their reproductive years. Surgical management can be accomplished hysteroscopically, laparoscopically, or abdominally depending on the size and location of the uterine fibroid.Large and/or multiple fibroids usually require an open approach. Transverse skin incisions are associated with less pain, quicker recovery, decreased hernia formation, and improved cosmesis. A Maylard incision allows for increased access to the pelvic sidewalls and can eliminate the need for a midline incision in patients with large pelvic pathology.There are no comprehensive, reliable educational videos for teaching this procedure. Curated instructional videos for surgical procedures can teach key surgical steps without needing to enter the operating room.This video is an excerpt from a full length teaching video demonstrating the key steps important for performing an abdominal myomectomy through a Maylard incision including patient positioning, relevant anatomy, and surgical technique. This video combines real-time operating room footage with high fidelity 3D animation.