Poor Video Quality Research Articles

Bruxism treatment on Youtube: evaluating reliability and information accuracy

BackgroundThe aim of this study was to evaluate the content and quality of videos about bruxism treatments on YouTube, a platform frequently used by patients today to obtain information.MethodsA YouTube search was performed using the keywords “bruxism treatment” and “teeth grinding treatment”. “The sort by relevance” filter was used for both search terms and the first 150 videos were saved. A total of 139 videos that met the study criteria were included in the study. Videos were classified as poor, moderate or excellent based on a usefulness score that evaluated content quality. The modified DISCERN tool was also used to evaluate video quality. Additionally, videos were categorized according to the upload source, target audience and video type. The types of treatments mentioned in the videos and the demographic data of the videos were recorded.ResultsAccording to the usefulness score, 59% of the videos were poor-quality, 36.7% were moderate-quality and 4.3% were excellent-quality. Moderate-quality videos had a higher interaction index than excellent-quality videos (p = 0.039). The video duration of excellent-quality videos was longer than that of moderate and poor-quality videos (p = 0.024, p = 0.002). Videos with poor-quality content were found to have significantly lower DISCERN scores than videos with moderate (p < 0.001) and excellent-quality content (p = 0.008). Additionally, there was a significantly positive and moderate (r = 0.446) relationship between DISCERN scores and content usefulness scores (p < 0.001). There was only a weak positive correlation between DISCERN scores and video length (r = 0.359; p < 0.001). The videos uploaded by physiotherapists had significantly higher views per day and viewing rate than videos uploaded by medical doctors (p = 0.037), university-hospital-institute (p = 0.024) and dentists (p = 0.006). The videos uploaded by physiotherapists had notably higher number of likes and number of comments than videos uploaded by medical doctors (p = 0.023; p = 0.009, respectively), university-hospital-institute (p = 0.003; p = 0.008, respectively) and dentists (p = 0.002; p = 0.002, respectively).ConclusionsAlthough the majority of videos on YouTube about bruxism treatments are produced by professionals, most of the videos contain limited information, which may lead patients to debate treatment methods. Health professionals should warn patients against this potentially misleading content and direct them to reliable sources.

Methods of video quality-improving

Video content has become integral to our daily lives, but poor video quality can significantly reduce viewers' experience and engagement. Various super-resolution methods are used to correct this, thereby reconstructing high-resolution videos from low-resolution ones. Two main categories of super-resolution methods exist traditional image processing and deep learning-based techniques. Deep learning-based techniques, such as Convolutional Neural Networks (CNNs), Generative Adversarial Networks (GANs), and Recurrent Neural Networks (RNNs), have shown great promise in enhancing video quality. The article discusses multiple adaptations of contemporary deep learning models to enhance video resolution. It also briefly explains the framework's design and implementation aspects. Lastly, the paper presents an overview and comparative analysis of the VSR techniques' efficiency on various benchmark datasets. At the same time, the paper describes potential challenges when choosing training sets; performance metrics, which can be used to compare different algorithms quantitatively. This work does not describe absolutely all existing VSR methods, but it is expected to contribute to the development of recent research in this field and potentially deepen our understanding of deep learning-based VSR methods, as well as stimulate further research in this area. In this work, new solutions for improving the performance of the methods are proposed, in particular, new quality metrics and datasets for model training. Overall, AI-based methods for VSR are becoming increasingly crucial with the rising demand for high-quality video content

Online Videos Regarding Relevant Postoperative Patient Information and Postoperative Rehabilitation After Arthroscopic Rotator Cuff Repair Provide Poor Information Quality, Accuracy, and Reliability

To evaluate the information quality, accuracy, and reliability of YouTube videos regarding relevant postoperative patient information and postoperative rehabilitation after arthroscopic rotator cuff repair. By use of The Onion Router (TOR) software and predefined search terms, 102 videos were assessed. Four scoring systems were used to evaluate included videos: (1) Journal of the American Medical Association (JAMA) benchmark criteria score; (2) Global Quality Score (GQS); (3) DISCERN score; and (4) a newly developed score, the Rotator Cuff Score (RCS). The RCS (0-30 points) was built based on the latest published evidence and guidelines from the American Academy of Orthopaedic Surgeons. Videos that scored up to 9 points were regarded as poor-quality videos. Most of the included videos provided poor information quality, accuracy, and reliability. Videos that were uploaded by medically trained professionals showed significantly better results for all scores compared with commercial or personal-testimony videos (JAMA benchmark criteria score, P < .001; GQS, P < .001; DISCERN score, P= .001; and RCS, P= .001). Multivariate linear regression showed that the involvement of medically trained professionals was a significant predictor of better results for all scores (JAMA benchmark criteria score, β= 1.496 [P < .001]; GQS, β= 1.105 [P < .001]; DISCERN score, β= 11.234 [P < .001]; and RCS, β= 5.017 [P < .001]). Surprisingly, the like ratio was significantly higher for videos that were uploaded by non-medically trained individuals (P= .041). The average information quality, accuracy, and reliability of YouTube videos regarding relevant postoperative patient information and postoperative rehabilitation after arthroscopic rotator cuff repair are poor. Videos from medically trained professionals provide significantly higher information quality; however, even these videos lack important information for a better understanding of arthroscopic rotator cuff repair. Because of the lack of a peer-review process, available videos on YouTube regarding relevant postoperative patient information and postoperative rehabilitation after arthroscopic rotator cuff repair are of low quality, accuracy, and reliability. However, patients increasingly visit YouTube to gather medical knowledge. Physicians should enlighten patients about these findings and should be able to provide alternative sources of high-quality information.

Assessment of YouTube as an information source for immune checkpoint inhibitors (ICIs).

11009 Background: YouTube is an open-access source of information. With more than 2.5 billion users, it can play a pivotal role in educating or misinforming cancer patients. ICI has changed the treatment paradigm for cancer over the past decade. There is scarce data to evaluate the quality of information on ICI uploaded on YouTube. We aim to assess the quality and engagement parameters of the 100 most viewed videos on YouTube focusing on ICI. Methods: A systematic search was conducted on YouTube using the search terms (“Immunotherapy AND Cancer”; “Immune Checkpoint Inhibitors AND Cancer”). The 100 most viewed English videos were selected. We utilized the already validated PEMAT1 and DISCERN2 tools to assess the quality of these videos on ICI. Additionally, we looked at the number of views, comments, and likes. Descriptive statistics and Pearson correlation tests were performed using R studio v2022.12.0-353. Results: The videos in our study had up to 42 million views (median views: 17,617). These videos were categorized by sources; academic (54%), media (31%), and industry (15%). The mean DISCERN score for the videos was 43.4. Only 22% videos were in the good or excellent category. Most videos (53%) discussed the benefits of ICI, whereas only 29% discussed the adverse effects of ICI. Shared decision-making was promoted by 34% of videos, and only 6.7% of videos summarized the information at the end, based on the DISCERN criteria. The medical terminologies were defined by 48.5% of videos, and 31% were broken into short sections with informative headers per PEMAT understandability criteria. There was no significant association between the video source and video engagement parameters, including likes, views, and comments. However, a trend towards an increased number of likes (r=-0.195, p=0.04) and comments (r=-0.213, p=0.03) was noticed with lower DISCERN scores (poor quality videos). Conclusions: The majority of YouTube videos are focused on the benefits of ICI. They are less likely to discuss the adverse effects or shared decision-making . In summary, YouTube videos contain biased information on ICI. Further efforts from academic institutions are needed to improve the quality and engagement of these videos. 1. Shoemaker SJ et al. The Patient Education Materials Assessment Tool (PEMAT) and User’s Guide; 2. Charnock D et al. DISCERN: an instrument for judging the quality of written consumer health information on treatment choices. J Epidemiol Community Health. 1999. Pearson correlation coefficient (r) between quality assessment scores and video variables. [Table: see text]

Telemedicine and YouTube™: Video quality analysis before and after COVID-19 pandemic

To assess the quality content of YouTube™ videos on telemedicine during COVID-19 pandemic. First, the frequency of worldwide YouTube™ and Google™ searches for telemedicine was analyzed. Second, we queried YouTube™ with telemedicine-related terms. Third, the Patient Education Materials Assessment Tool for Audiovisual Materials (PEMAT A/V), the Global Quality Score (GQS), and the Misinformation tool were used for the quality assessment. According to selection criteria, 129 videos were collected for the analysis. From January 2018 to January 2022, the peak relative interest on YouTube™ and Google™ occurred in March 2020. Of all, 27.1 and 72.9% were uploaded before (Jan 2018-Feb 2020) and after (Mar 2020-Mar 2022) the COVID-19 outbreak, respectively. According to the PEMAT A/V, the overall median understandability and actionability was 50.0% (33.3 [IQR 0-66.7] vs 50.0 [27.1-75], p = 0.2) and 66.7% (63.6 [IQR 50.0-75.7] vs 67.9 [50.0-79.2],p = 0.6), respectively. According to GQS, 3.9%, 17.8%, 24.0%, 26.4% and 27.9% were classified as excellent, good, medium, generally poor, and poor-quality videos, respectively. The highest rate of poor-quality videos was recorded in videos uploaded before COVID-19 pandemic (37.1 vs 24.5%). According to overall misinformation score, a higher score was recorded for the videos uploaded after COVID-19 pandemic (1.8 [IQR 1.4-2.3] vs 2.2 [1.8-2.8], p = 0.01). The interest in telemedicine showed a significant peak when the COVID-19 pandemic was declared. However, the contents provided on YouTubeTM were not informative enough. In the future, official medical institutions should standardize telemedicine regulation and online content to reduce the widespread of misleading information.

Longitudinal analysis of YouTube videos discussing post-prostatectomy erectile dysfunction.

YouTube is heavily utilized by patients as an educational resource, but this content can be fraught with misinformation. We sought to characterize the quality of videos on YouTube discussing postprostatectomy erectile dysfunction and to evaluate metrics associated with retaining a top position in search results over time. In October 2019, we watched the first 100 YouTube videos using the search query "radical prostatectomy erectile dysfunction." Videos not relevant to the topic were excluded. Video metrics were collected, and content quality was evaluated using the DISCERN instrument. In June 2022, the search was repeated and video metrics were updated. Video characteristics were associated with search rank and the ability to remain in the top 100 spots using the Pearson correlation coefficient (r) and logistic regression, respectively. We included 81 videos which amassed 529,428 views in 2019. The median total DISCERN score was 29 (IQR 21-42), which is interpreted as a poor quality video. Self-promotion or commercial bias was present in 42 videos (51.9%); false claims were present in 16 (19.8%). There was no correlation between DISCERN score and search rank (r = 0.08, p = 0.49). In 2022, 15 videos remained in the top 100 search results and had a higher median DISCERN score than videos no longer in the top 100 (46 vs. 28.5, p = 0.01). Each additional DISCERN point was associated with a 7% higher odds of remaining in the top 100 (OR 1.07, 95% CI 1.01-1.11, p = 0.003). The quality of the top 100 YouTube videos discussing postprostatectomy erectile dysfunction is low. Higher quality videos had a higher odds of remaining in the top 100 search results over time but do not correlate with the order in which they are ranked.

YouTube™ as a source of information on prostatitis: a quality and reliability analysis.

It was aimed to analyze the YouTube™ videos on prostatitis regarding their source, content, and information included. The term "prostatitis" was searched by relevance and the first 200 video links and features were recorded. Using the 5-point modified DISCERN tool, Global Quality Score (GQS), and Journal of American Medical Association (JAMA) score, the quality and reliability of the information were assessed by two urologists. Inter-rater agreement for DISCERN, JAMA, and GQS had Cohen's kappa coefficients of 0.883, 0.887, and 0.885, respectively. The most common source of upload was doctors/medical institutions (33.0%), and the majority of the content was about general information (27.0%). The mean modified DISCERN, GQS and JAMA scores were 3.24 ± 1.23 (1.00-5.00), 3.25 ± 1.26 (1.00-5.00) and 2.17 ± 1.36 (0.00-4.00), respectively. These scores were correlated with each other (r = 0.914, p < 0.001 between JAMA and GQS; r = 0.954, p < 0.001 between JAMA and modified DISCERN; and r = 0.885, p < 0.001 between GQS and modified DISCERN, Spearman test) and were the highest in the videos uploaded by doctors/medical institutions (p < 0.001, Kruskal-Wallis test). Of the videos, 25.5% were poor, 30.5% were moderate and 44.0% were high quality. The high quality videos were significantly shorter than those with poor quality (p = 0.039, Kruskal-Wallis test). The poor quality videos had the highest view ratio (50.49 ± 127.74 (0.03-618.91), p = 0.036, Kruskal-Wallis test). Most YouTube™ videos on prostatitis are short and high quality videos uploaded by doctors/medical institutions. However, these videos have less view ratios than those with poor quality.

Noise quality estimation on portraits in realistic controlled scenarios

The wide use of cameras by the public has raised the interest of image quality evaluation and ranking. Current cameras embed complex processing pipelines that adapt strongly to the scene content by implementing, for instance, advanced noise reduction or local adjustment on faces. However, current methods of Image Quality assessment are based on static geometric charts which are not representative of the common camera usage that targets mostly portraits. Moreover, on non-synthetic content most relevant features such as detail preservation or noisiness are often un-tractable. To overcome this situation, we propose to mix classical measurements and Machine learning based methods: we reproduce realistic content triggering this complex processing pipelines in controlled conditions in the lab which allows for rigorous quality assessment. Then, ML based methods can reproduce perceptual quality annotated previously. In this paper, we focus on noise quality evaluation and test on two different set ups: closeup and distant portraits. These setups provide scene capture conditions flexibility, but most of all, allow the evaluation of all quality camera ranges from high quality DSLR to poor quality video conference. Our numerical results show the relevance of our solution compared to geometric charts and the importance of adapting to realistic content.

Teachers’ Perceptions of the Platforms Used for Assessing Learners’ Speaking Performance: Types, Benefits, and Challenges

In the teaching and learning process, one of the roles and responsibilities of teachers is to assess the skills and what is taught to students. In the assessment activity, the teacher can get feedback on the extent to which students understand what the teacher teaches. Even though the learning is done virtually, the assessment still has to be done. This study investigates teachers’ perceptions of the types, benefits, and challenges of platforms used to assess young learners’ speaking performance. The research data was obtained by conducting interviews with respondents through Google Meet and adopted the interview questions by Mpungose (2020) about the online assessment platforms. The researcher determined the respondents by using a purposive sampling technique where respondents used an online platform in assessing young learners’ speaking performance during distance learning. After the data was collected, the researcher analyzed it using an interactive model by Miles et al. (2014). This study indicates that there are various platforms that teachers can use to assess young learners’ speaking performance, including Zoom, Google Meet, Google Classroom, WhatsApp, and YouTube. The five platforms were revealed to be very useful to make it easier for teachers to assess learners’ speaking performance. Teachers can easily assess aspects of their pronunciation, fluency, grammar, and vocabulary through these five platforms. Besides helping assess learners’ speaking performance, this study also shows various challenges in its implementation, such as connection problems, assignment submission, and poor video and audio quality.

Evaluation of an online otorhinolaryngology clinical training module for undergraduate medical students implemented during COVID 19 pandemic

The COVID 19 has posed many challenges to the health care educators. The outbreak made health educators to introduce and integrate information technology into the medical teaching. In the endeavour to create a virtual clinical learning environment, for the subject, the Dept. of Otorhinolaryngology, ELMCH structured an online clinical training module. The study was undertaken to implement and assess the effectiveness of the module. The module was divided into parts covering various aspects of the clinical rotation. After completion of the clinical posting all students, faculty and residents were asked to fill a pre-validated questionnaire on google form to assess the effectiveness and their perceptions for online clinical teaching module. Most of the students felt that the session met their learning needs (4.4±.87) and interactiveness was good(4.5±.72). Most of the students also felt that the session were convenient (4.4±.84) and helped them in gaining clinical skills (4.3±.87). The most common difficulty in e-learning was technical problems including unstable internet connectivity resulting in poor voice and video quality.

Video Super-Resolution for Wireless Capsule Endoscopy Imaging Sensor

As a preferred diagnostic imaging sensor for small bowel diseases, wireless capsule endoscopy (WCE) suffers from poor video quality, which can be mitigated via video super-resolution (VSR) techniques. However, the promise of applying VSR into WCE applications has been offset by two significant challenges. Firstly, high-resolution (HR) counterparts of real endoscopic frames are unavailable for supervised learning. Secondly, utilizing temporal correlations to enhance recovery performance becomes difficult due to the poor video quality. To tackle these two challenges, we develop a novel training dataset and an effective alignment network for VSR on practical WCE applications. Specifically, clean natural videos and synthetic endoscopic videos are collected as the ground truth of the training dataset. Then, a set of complex degradation models is designed to generate their low-resolution (LR) counterparts for supervised learning. Although real HR endoscopic videos remain unknown during training, this complicated dataset enables the trained network to have a strong generalization capacity so that it can make reasonable inferences on real endoscopic videos. To utilize temporal correlations from WCE videos, a block-based temporal attention alignment network (BTAAN) is proposed to build accurate and robust correspondences among consecutive frames under poor video quality. This network exploits the global information from neighboring frames to capture large-scale motions, and suppresses the effect of unreliable points by a temporal attention mechanism. Moreover, its block-wise characteristic reduces the computational complexity significantly, which makes it hardware-friendly. Extensive experiments demonstrate the superiority of the proposed method over other state-of-the-art approaches on practical WCE applications.

Injury mechanism of knee medial collateral ligament: An online systematic video analysis

Objectives: The medial collateral ligament (MCL), one of the main stabilizers of the knee, is also the most injured ligament of the knee. Isolated MCL injuries are common in young men who play contact sports. It was hypothesized that MCL rupture in professional athletes develops with a specific injury mechanism. Furthermore, body posture at the time of injury varies in different sports. Materials and Methods: On May 15, 2021, “knee medial collateral injury” and “MCL injury” were entered in a YouTube® search. The inclusion criteria were to verify that the video was compatible with the injury, that the video included the foot, ankle, knee, hip, and the whole body at the time of injury, and that image quality was sufficient to detect the mentioned body parts. Exclusion criteria were non-MCL knee injuries, insufficient images, poor video quality, and additional injuries with MCL. Results: Of the 23 injuries included in the study, 22 (95.7%) were male and 1 was female (4.3%). The distribution of activities in which injuries occurred was seven basketball (30.4%), four American football (17.4%), four ice hockey (17.4%), three soccer (13.0%) one water ski (4.3%), one skateboard (4.3%), two-track and field (8.6%), and one wrestling (4.3%). In 18 cases (78.3%), the injury occurred by contact. Of these, 16 were caused by direct contact (69.6%) and two by indirect contact (8.7%). Conclusion: Isolated MCL injury occurs when valgus force is applied with the trunk flexed and turned towards the injured side, the hip abducted and slightly flexed, and the knee slightly flexed. The majority of these cases occur after a direct blow to the lateral knee.

Dynamic GAN for high-quality sign language video generation from skeletal poses using generative adversarial networks

The recent advancements of unsupervised deep generative models have produced incredible results in image and video generation tasks. However, existing approaches still pose huge challenges in the high-quality video generation process. The generated videos consist of blurred effects and poor video quality. In this paper, we introduce a novel generative framework named dynamic generative adversarial networks (dynamic GAN) model for regulating the adversarial training and generating photo-realistic high-quality sign language videos. The proposed model uses skeletal poses information and person images as input and produces high-quality videos. In generator phase, the proposed model uses U-Net-like network to generate target frames from skeletal poses. Further, the generated samples are classified using the VGG-19 framework to identify its word class. The discriminator network classifies the real and fake samples as well as concatenates the resultant frames and generates the high-quality video output. Unlike, existing approaches the proposed novel framework produces photo-realistic video quality results without employing any animation or avatar approaches. To evaluate the model performance qualitatively and quantitatively, the proposed model has been evaluated using three benchmark datasets that yield plausible results. The datasets are RWTH-PHOENIX-Weather 2014T dataset, and our self-created dataset for Indian Sign Language (ISL-CSLTR), and the UCF-101 Action Recognition dataset. The proposed model achieves average 28.7167 PSNR score, 0.921 average SSIM score, 14 average FID score and 8.73 ± 0.23 average inception score which are relatively higher than existing approaches.

How Physicians in a Moroccan Tertiary Care Center Perceive Teleconsultation During COVID-19 Pandemic?

Background: Recent technological advances have paved the way for a new modality of medical practice known as teleconsultation. Positive perceptions about the benefits of teleconsultation and its acceptance by clinicians are key predictors of its uptake. The aim of this study was to assess knowledge, perceptions, and acceptability of teleconsultation among Moroccan physicians. Methods: This is a descriptive cross-sectional study conducted at the Cheikh Khalifa International University Hospital (HCK) of Casablanca, Morocco. Study participants answered an anonymous self-administered questionnaire designed based on similar studies. The questionnaire contained four sections related to demographic characteristics, knowledge about telemedicine and teleconsultation, perceptions about teleconsultation, and its acceptability by the study participants. Results: Of 486 eligible participants, a total of 212 completed the questionnaire. Approximately 96.7% had prior knowledge of teleconsultation. Most participants identified internet access (95.3%), data security (93.4%), and confidentiality (92.9%) as the main facilitators to the use of teleconsultation. The main barriers to the use of teleconsultation were internet access issues (98.6%), poor audio quality (96.2%), poor video quality (94.3%), and difficulty encountered by patients in expressing themselves and communicating with their physician (79.7%). The majority of participants (91.5%) believed that teleconsultation will be an integral part of future medical practice and about 70.8% thought that they would be able to allot time to teleconsultation in their current schedules. Conclusions: The findings of this study should be used by policy makers to remove barriers and promote enablers of teleconsultation use by physicians to bring health care closer to the Moroccan population.

YouTube-videos for patient education in lymphangioleiomyomatosis?

BackgroundThe Internet is commonly used by patients to acquire health information. To date, no studies have evaluated the quality of information available on YouTube regarding lymphangioleiomyomatosis (LAM). Our aim was to determine the quality and content of YouTube videos regarding LAM and to compare the information provided with current knowledge and guidelines about the disease.MethodsThe first 200 video hits on YouTube in English for the search term “lymphangioleiomyomatosis” were recorded. All videos suitable for patient education on LAM were included. Video quality was analyzed independently by two investigators utilizing the Health on the Net (HONcode) score, which assesses whether websites provide understandable, accessible, and trustworthy health information; the DISCERN score, which evaluates the quality of information about treatment decisions; and a newly developed LAM-related content score (LRCS) with 31 guideline elements.ResultsThe search identified 64 eligible videos. The “engagement rate” of 0.3 was low, with a median number of views of 408 (range 42–73,943), a median of 4 likes (range 0–2082), and the majority (53%) receiving a low HONcode score (≤ 2) and only 10% of videos achieving a high score (> 5). The median DISCERN score was 28 (range 15–61, maximum possible score 80), indicating poor video quality and reliability. The median LRCS was 8 (range 0–29, maximum possible score 31) and videos frequently failed to provide sources of information.ConclusionsOnline resources could contribute to the limited and often inaccurate information available to patients with LAM, with only a few YouTube videos providing high-quality patient-relevant information.

Usage Characteristics of Mobile Telecommunication Services: Individual Service Recipient Perspective in Rajshahi District of Bangladesh

Mobile telecommunication services proved to be a significant part of everyday life for the people of Bangladesh. The study's primary purpose was to examine the usage characteristics of subscribers to know the reality through an empirical study. The research study was descriptive and quantitative in approach to drawing meaningful facts. This study was conducted in the Rajshahi district in the northern region of Bangladesh. The sample size resulted in three hundred and seventy-eight for completing the survey. A simple random sampling technique was used to draw samples from male and female sampling units. The research used a cross-sectional survey design to collect primary data. In addition, the researcher administered a structured questionnaire as a research tool to find out the facts from the individual service recipients. Research findings show that individual service recipients’ usage characteristics influence the service marketing mix and brand equity to achieve a competitive advantage. For example, the study found that consumers are more interested in using services other than voice calls through their affordability does not support it. Moreover, responses from service recipients varied significantly, which becomes a significant challenge for the operators to blend the elements of the marketing mix. This study also shed light on the Inconvenience or problems the recipients face in consuming services. For example, the topmost inconvenience service recipients face is 'poor service quality’ like network interruption, poor quality network, and poor quality video streaming experience. Finally, the study suggests mobile operators in Bangladesh need to trigger a marketing mix by consistently considering the dynamism of usage characteristics of individual service recipients to utilize their marketing investment, increase brand equity, and achieve a significant competitive advantage.

Does social media provide adequate health education for prevention of COVID-19? A case study of YouTube videos on social distancing and hand-washing.

Social media offers an opportune platform for educating the public about the recommended interventions during global health emergencies. This case study evaluated information in the popular social media platform YouTube about two key interventions (namely, ‘social distancing’ and ‘hand washing’) recommended during coronavirus disease-2019. Using the keywords ‘social distancing’ and ‘hand washing’, 77 and 78 videos, respectively, were selected from YouTube through pre-defined criteria. The understandability, actionability and quality of information in these videos were assessed. Cumulatively, the social distancing videos received >9 million views and the hand-washing videos received >37 million views. Thirteen social distancing videos (16.9%) and 46 hand-washing videos (58.9%) provided understandable, actionable and good-quality information. The non-understandable, non-actionable or poor-quality videos had paradoxically more viewer engagements than the understandable, actionable or good-quality videos, respectively. Most social distancing videos came from news agencies (68.8%). Hand-washing videos were mostly uploaded by health agencies or academic institutes (52.6%). The videos were less likely to be understandable and actionable and to be of good quality when uploaded by sources other than health agencies or academic institutes. The paucity of adequate information and the limited representation of ‘authoritative’ sources were concerning. Strategies for harnessing social media as an effective medium for public health education are necessary during pandemics.

OldVSR: A model for the video super-resolution and restoration of old real-world TV series

With the recent advance in video super-resolution (VSR) techniques, there have been many requests for super-resolve real-world old analog TV series into high-definition digital content. As excellent classical TV series may receive little to no attention due to their poor video quality, restoring them would open new business opportunities for reusing old TV contents. A problem with restoring real-world old TV series is in the complex artifacts introduced by the old interlaced scanning and compression artifacts during the digitization of old analog videos. Though recent DNN-based VSR models perform nicely on clean videos, due to the artificial nature of interlacing and compression artifacts, they fail to restore old videos into a high-definition counterpart free from noticeable artifacts. In this work, we propose OldVSR for restoring old real-world TV series with artifacts of artificial nature. The proposed model implements a bidirectional recurrent structure with first and second-order propagation where each recurrent layer implements two main functions, i.e., Feature alignment (FA) and Pyramid feature aggregation (PFA). The outputs of the forward and backward layers are merged and upsampled to produce a High-Definition (HD) frame of the input standard-definition (SD) frame. We demonstrate through experiments that our proposed OldVSR can effectively remove artifacts of artificial nature from old videos and successfully restores old TV series.

Correcting Face Distortion in Wide-Angle Videos.

Video blogs and selfies are popular social media formats, which are often captured by wide-angle cameras to show human subjects and expanded background. Unfortunately, due to perspective projection, faces near corners and edges exhibit apparent distortions that stretch and squish the facial features, resulting in poor video quality. In this work, we present a video warping algorithm to correct these distortions. Our key idea is to apply stereographic projection locally on the facial regions. We formulate a mesh warp problem using spatial-temporal energy minimization and minimize background deformation using a line-preservation term to maintain the straight edges in the background. To address temporal coherency, we constrain the temporal smoothness on the warping meshes and facial trajectories through the latent variables. For performance evaluation, we develop a wide-angle video dataset with a wide range of focal lengths. The user study shows that 83.9% of users prefer our algorithm over other alternatives based on perspective projection. The video results can be found at https://www.wslai.net/publications/video_face_correction/.

Standardization of ileoanal J-pouch surgery technique: Quality assessment of minimally invasive ileoanal J-pouch surgery videos

BackgroundIleal pouch anal anastomosis is a complex procedure associated with significant morbidity, with several complications after ileal pouch anal anastomosis surgery leading to pouch failure. The aim of the study is to evaluate the heterogeneity surrounding the technique of ileoanal J-pouch surgery by assessing the safety and quality of published online peer-reviewed surgical videos. MethodsIleal pouch anal anastomosis videos published on peer-reviewed surgical journals and video channels were edited and anonymized to demonstrate specific steps of the surgical procedure: mobilization and division of the rectum, formation of the ileoanal J-pouch reservoir, anastomosis, and lengthening techniques. The anonymized videos were presented to a group of reviewers with expertise in ileal pouch anal anastomosis blinded to the names and affiliations of the surgeons performing the procedure. Primary outcome was the rate of interobserver variability in the assessment of specific technical steps of the ileal pouch anal anastomosis surgery procedure. Secondary outcome was the appropriateness of the use of surgical videos review as an assessment tool for ileal pouch anal anastomosis surgery, measured as rate of reviewers being unable to answer for poor video quality. ResultsIn total, 29 video fragments were distributed, and 13 assessors completed a 60-item survey, organized in 7 major domains. The survey completion rate was 93.4%. Out of a total 729 answers, in 23 (3.2%) the reviewers indicated they were unable to comment due to poor video image, and in 48 (6.5%) were unable to comment due to the particular step not being shown in the procedure. The proportion of assessors rating rectal mobilization technically appropriate ranged from 30.7% to 92.3% and from 7.7% to 69.2% for safety. The level of rectal division was considered appropriate in 0 to 53.8% of the videos, whereas the stapling technique used for rectal division was appropriate in 0 to 70% of the videos. ConclusionOur study assessed published peer-reviewed videos on ileal pouch anal anastomosis surgery and reported heterogeneity in the safety of the demonstrated techniques. Blind assessment of published peer-reviewed ileal pouch anal anastomosis videos reported a high rate of unsafe or inappropriate technique for rectal mobilization and transection in the reviewed videos, with fair interobserver agreement among reviewers. There is a need for consensus on what is considered safe and appropriate in ileal pouch anal anastomosis surgery. Peer review of ileal pouch anal anastomosis surgery videos could facilitate training and accreditation in this complex procedure.