Individual Video Research Articles

Shared Brains, Proprioceptiveness, and Critically Approaching the Animal as the Animal in Artworks

The animal and being animal is a proposition and position that invites observational and critical debate. Yet, the presence of the non-human animal is usually and normatively confined to representational artworks rather than the animal itself in the gallery or museum, which is, potentially, problematically anthropocentric. Using diverse methods, processes, and materials, and curious to a myriad of opening potentialities, Bartram + Deigaard, in contrast to this problem, explore working as humans from an animal-centric perspective through artistic research. They bring sensitivities to their handling of the animal, as both artistic subject and collaborator, to observe and engage with empathy and openness to animal insight and revelation and behaviour. Their works in performance, video, drawing, and printmaking foreground animal proximity and behaviour, inter-species proprioception, reciprocal caretaking, synchronised respiration, and companionate movement. This article explores the socialised and familiar in close observation, directly and indirectly, in their individual yet companion practices, illuminating the benefits of a radically enlarged sentiocentrism. It reflects on the allowing and embracing of other species within their artworks, and of being mindful and sensible with balancing sympathies and empathies as humans within an often unbalanced system of agency. Specifically, it gleans patterns and insights from their exhibition at Tippetts and Eccles Galleries at Utah State University in 2021, where they invited a canine collaborator into their thinking through praxis and the interventions and residual outcomes this created. This essay discusses two individual video artworks from each artist, which document their invitations to non-human animals into the gallery or museum, and two durational artworks curated within this exhibition.

A method supporting fault-tolerant optical text recognition from video sequences recorded with handheld cameras

In the paper a method supporting the optical character recognition from video sequences recorded with cameras without good stabilization is proposed. Due to the presence of various distortions, such as motion blur, shadows, lossy compression artifacts, auto-focusing errors, etc., the quality of individual video frames, e.g., recorded by a smartphone camera, differs noticeably, influencing the results of text recognition, causing the presence of additional errors that may be unacceptable even in fault-tolerant applications. Although the quality of individual video frames may be assessed using state-of-the-art no-reference image quality metrics, the obtained results, particularly using the relatively fast metrics, are poorly correlated with recognition results. Therefore, a novel hybrid approach to image quality assessment supporting fault-tolerant optical character recognition is proposed, making it possible to select automatically the video frames leading to the smallest number of errors. Since the data extraction is rarely perfect, the proposed method may be particularly useful in systems where users should identify the faults defined as a too high number of text recognition errors. The proposed method makes it possible to eliminate low quality fragments of video sequences from further analysis, reducing the overall computational burden. It may also be useful in industrial OCR applications based on offline video recording with memory constraints where the use of online OCR services is not possible.

How to increase cookstove adoption? Exploring cost-effective dissemination techniques in Central Mozambique

The adoption and usage of improved cookstoves have been promoted in the Global South to mitigate the adverse effects of relying on solid biomass for cooking. However, the uptake and continued use of improved cookstoves remain low. Several behavioral changing techniques have been used to increase the adoption of improved cooking stoves. Still, the monetary costs involved in the dissemination process are generally neglected, and few studies follow an experimental design to evaluate the impact of different stove dissemination approaches. Thus, we conducted an experiment to explore the cost-effectiveness of dissemination approaches and measure their effect on improved cookstoves adoption in Central Mozambique. Forty communities were randomly assigned to receive one of the following treatments: (1) individual video training, (2) on-site group training or (3) a combination of both. In total, 620 households we randomly selected from the 40 communities to participate in the study and 510 were included in the final analysis. The rates of adoption within each treatment were 53 % for individual video training, 21 % for on-site group training and 47 % for the combination of both. The monetary training costs were lower for individual video training, which stimulated the early and sustained adoption of improved cook stoves. Our findings provide strong evidence that the tool used for capacity building affects the adoption of improved cookstoves. Extension agencies should be aware of cost-effectiveness and adjust dissemination approaches to achieve the highest impact per unit of resource expenditure.

Werewolf-XL: A Database for Identifying Spontaneous Affect in Large Competitive Group Interactions

Affective computing and natural human-computer interaction, which would be capable of interpreting and responding intelligently to the social cues of interaction in crowds, are more needed than ever as an individuals affective experience is often related to others in group activities. To develop the next-generation intelligent interactive systems, we require numerous human facial expressions with accurate annotations. However, existing databases usually consider nonspontaneous human behavior (posed or induced), individual or dyadic setting, and a single type of emotion annotation. To address this need, we created the Werewolf-XL database, which contains a total of 890 minutes of spontaneous audio-visual recordings of 129 subjects in a group interaction of nine individuals playing a conversational role-playing game called Werewolf. We provide 131,688 individual utterance-level video clips with internal self-assessment of 18 non-prototypical emotional categories and external assessment of pleasure, arousal, and dominance, including 14,632 speakers' samples and the rest of listeners' samples. Besides, the results of the annotation agreement analysis show fair reliability and validity. Role information and outcomes of the game are also recorded. Furthermore, we provided extensive benchmarks of unimodal and multimodal emotional recognition results. The database is made publicly available.

ICatcher+: Robust and Automated Annotation of Infants' and Young Children's Gaze Behavior From Videos Collected in Laboratory, Field, and Online Studies.

Technological advances in psychological research have enabled large-scale studies of human behavior and streamlined pipelines for automatic processing of data. However, studies of infants and children have not fully reaped these benefits because the behaviors of interest, such as gaze duration and direction, still have to be extracted from video through a laborious process of manual annotation, even when these data are collected online. Recent advances in computer vision raise the possibility of automated annotation of these video data. In this article, we built on a system for automatic gaze annotation in young children, iCatcher, by engineering improvements and then training and testing the system (referred to hereafter as iCatcher+) on three data sets with substantial video and participant variability (214 videos collected in U.S. lab and field sites, 143 videos collected in Senegal field sites, and 265 videos collected via webcams in homes; participant age range = 4 months-3.5 years). When trained on each of these data sets, iCatcher+ performed with near human-level accuracy on held-out videos on distinguishing "LEFT" versus "RIGHT" and "ON" versus "OFF" looking behavior across all data sets. This high performance was achieved at the level of individual frames, experimental trials, and study videos; held across participant demographics (e.g., age, race/ethnicity), participant behavior (e.g., movement, head position), and video characteristics (e.g., luminance); and generalized to a fourth, entirely held-out online data set. We close by discussing next steps required to fully automate the life cycle of online infant and child behavioral studies, representing a key step toward enabling robust and high-throughput developmental research.

Exploring Deepfakes - Creation Techniques, Detection Strategies, and Emerging Challenges: A Survey

Abstract: Deep learning, integrated with Artificial Intelligence algorithms, has brought about numerous beneficial practical technologies. However, it also brings up a problem that the world is facing today. Despite its innumerable suitable applications, it poses a danger to public personal privacy, democracy, and corporate credibility. One such use that has emerged is deepfake, which has caused chaos on the internet. Deepfake manipulates an individual's image and video, creating problems in differentiating the original from the fake. This requires a solution in today's period to counter and automatically detect such media. This study aims to explore the techniques for deepfake creation and detection, using various methods for algorithm analysis and image analysis to find the root of deepfake creation. This study examines image, audio, and ML algorithms to extract a possible sign to analyze deepfake. The research compares the performance of these methods in detecting deepfakes generated using different techniques and datasets. As deepfake is a rapidly evolving technology, we need avant-garde techniques to counter and detect its presence accurately.

A vision transformer for decoding surgeon activity from surgical videos

The intraoperative activity of a surgeon has substantial impact on postoperative outcomes. However, for most surgical procedures, the details of intraoperative surgical actions, which can vary widely, are not well understood. Here we report a machine learning system leveraging a vision transformer and supervised contrastive learning for the decoding of elements of intraoperative surgical activity from videos commonly collected during robotic surgeries. The system accurately identified surgical steps, actions performed by the surgeon, the quality of these actions and the relative contribution of individual video frames to the decoding of the actions. Through extensive testing on data from three different hospitals located in two different continents, we show that the system generalizes across videos, surgeons, hospitals and surgical procedures, and that it can provide information on surgical gestures and skills from unannotated videos. Decoding intraoperative activity via accurate machine learning systems could be used to provide surgeons with feedback on their operating skills, and may allow for the identification of optimal surgical behaviour and for the study of relationships between intraoperative factors and postoperative outcomes.

Assessment of Video Quality and Reliability of YouTube Videos Regarding Meniscus Tear Rehabilitation.

YouTube is an open-source platform where creators can record and upload videos for others to see. As the popularity of YouTube increases, it is being increasingly used for healthcare-related information. However, with the relative ease of uploading videos, the content quality of individual videos is not regulated. This study aimed to assess and analyze the content quality of YouTube videos concerning meniscus tear rehabilitation. We hypothesized that most videos would be of low quality. The keywords "meniscus tear treatment," "meniscus tear recovery," "meniscus tear physical therapy," and "meniscus tear rehabilitation" were used for searching videos on YouTube. A total of 50 videos was analyzed in this study that related to meniscal rehabilitation, grouped into one of four categories: non-physician professionals (physical therapists and chiropractors) (n=28), physicians (with or without academic affiliation) (n=5), non-academic healthcare-related websites (n=10), and non-professional individuals (n=7). Videos were then assessed by two independent authors using three scoring systems, which included the Global Quality Scale (GQS), modified DISCERN, and Journal of the American Medical Association (JAMA) scores. The number of likes, comments, video length, and views were collected for each video. Kruskal-Wallis tests were used to compare quality scores and video analytics. The overall median GQS, modified DISCERN, and JAMA scores were 3 (interquartile range (IQR): 2-3), 2 (IQR: 2-2), and 2 (IQR: 2-2), respectively. Sorted by GQS scores, 20 videos were of low quality (40%), 21 were intermediate (42%), and nine were of high quality (18%). Non-physician professionals produced 56% (28 of 50) of the videos assessed, with physical therapists constituting 86% (24 of 28) of this category. The median duration of each video was 6:54 (IQR: 3:59-10:50) minutes, views was 42,262 (IQR: 12,373-306,491), and likes was 877 (IQR: 239-4,850). Kruskal-Wallis testing showed a significant difference between JAMA scores, likes, and video length duration between video categories (p<0.028). The overall median reliability of YouTube videos on meniscus tear rehabilitation measured by JAMA and modified DISCERN scores was low. The overall median video quality, as assessed by GQS scores, was intermediate. Video quality was highly variable, with fewer than 20% meeting high-quality standards. As a result, patients are often viewing lower quality videos when researching their condition online.

It's Not Just What You Say, It's How You Say It: Mitigating the Impacts of Greenhouse Gas Emissions with Effective Content

Purpose: The goal of this article is to provide a tutorial for measuring the sustainability impacts of our content. Sustainability considerations should be just as important as the accessibility and usability of content. Fortunately, following our known best practices for effective content creation and governance also results in a smaller carbon footprint. This not only benefits the planet but also positions content creators as valuable business assets. Method: An extensive literature review led to the development of processes and guidelines that connect our existing best practices for content with a new paradigm for sustainable content. Results: This tutorial will examine five content scenarios and the relative impacts of each: modifications to individual web pages, content governance (removal of redundant, outdated, and trivial content), email blasts, videos, and podcasts. Metrics will be provided for justifying content decisions to management, and recommendations will be given for both optimizing usability and sustainability. Conclusion: Merging established best practices for content creation and curation with measurable energy-to-emissions data gives us powerful tools for justifying content choices and demonstrating measurable value to both the organization and the planet.

Adoption of video consultations during the COVID-19 pandemic

BackgroundVideo consultations have the potential to play a significant role for the future of healthcare by solving some of the imminently arising healthcare challenges, as pointed by the European Commission in Europe and the National Academy of Medicine in the United States of America. This technology can improve quality, efficiency, and enhance access to healthcare. ObjectiveThe aim of this study is to explore and understand individual video consultations acceptance drivers. MethodsAn extended technology acceptance model was created based on the diffusion of innovation theory (DOI), unified theory of acceptance and use of technology (UTAUT), health belief model (HBM), and concerns for information privacy framework (CFIP). 346 valid responses were collected through an online questionnaire, and the partial least squares (PLS) modeling approach was used to test the model. ResultsThe model explained 77.6 % (R2) of the variance on intention to use, and 71.4 % (R2) of the variance in attitude. The predictors of intention to use are attitude (beta = 0.504, p-value<0.001), performance expectancy (beta = 0.196, p-value = 0.002), and COVID-19 (beta = 0.151, p-value<0.001). The predictors of attitude are performance expectancy (beta = 0.643, p-value>0.001), effort expectancy (beta = 0.138, p-value = 0.001), and COVID-19 (beta = 0.170, p-value<0.001). ConclusionsThis research model highlights the importance of creating extended acceptance models to capture the specificities of each technology in healthcare. The model created helps to understand the most important drivers of video consultation acceptance, highlighting the importance of the COVID-19 pandemic and perceived health risks.

An analysis of a video education campaign on COVID-19 for healthcare workers in Africa delivered through Facebook.

The COVID-19 pandemic required rapid dissemination of accurate information across the world to both healthcare workers and the general public. Social media represents an opportunity to undertake this. The aim of this study was to analyse a healthcare worker education campaign in Africa delivered through the social media platform Facebook and discuss the feasibility of this approach for future healthcare workers and public health campaigns. The campaign ran from June 2020 to January 2021. The Facebook Ad Manager suite was used to extract data in July 2021. Videos were analysed for total and individual video reach, impressions, 3-second video plays, 50% plays and 100% plays. The geographic use of the videos and age and gender breakdown was also analysed. Total reach of the Facebook campaign was 6,356,846 and total impressions was 12,767,118. The video with the highest reach was 'Hand washing steps for health workers' with a reach of 1,479,603. The total campaign 3-second plays were 2,189,460 decreasing to 77,120 for 100% play duration. Facebook advertising campaigns may have the ability to reach large populations and achieve a range of engagement outcomes that would be more cost effective and have greater reach when compared with traditional media. The outcome of this campaign has shown the potential of social media's use in public health information, medical education and professional development.

How general practitioners perceive the aging trajectory of oldest-old - A qualitative study

BackgroundA new population of older people is growing: the oldest-old. The care of the oldest-old (individuals aged 90 and over) is a new challenge in primary care. This study aimed to analyze the perception of General Practitioners (GP) on (1) the aging process of their patients up to a very advanced age, (2) how to adapt their practice to the care of these patients.MethodsWe conducted a qualitative study using focus group (face to face) and individual (video call) interviews of GPs in southwest France. The sampling was purposive. We analyzed the interviews using an inductive approach based on the phases of thematic analysis. We used researchers’ triangulation during the process. Collection was concluded when saturation was reached.ResultsThree focus groups and one individual interview were conducted with a total of seventeen general practitioners. GP perception concerning aging and very advanced age were based on their personal experience and their daily clinical practice. Aging was perceived as an individual, unconscious, unpredictable and irreversible phenomenon. The shift towards “very old age” appeared inevitable. It could be a physical or psychological shift, or patients neglecting themselves or lacking a project. The care of the oldest-old became more specific and individual, adapted to the wishes of the patient. Those adaptations involve medical disengagement to focus on the most essential outcomes. The objectives of health care needed to be less strict with limited invasive practices. Prevention needed to focus mainly on prevention of falls and limitation of functional decline.ConclusionGPs identified an inevitable and unpredictable shift from old age to very old age. The adaption of the theory of disengagement allowed us to identify a medical disengagement of the GPs in the care of their oldest patients.

In-Field Pine Seedling Counting Using End-to-End Deep Learning for Inventory Management

Highlights A deep learning-based machine vision system was developed for pine seedling counting. Automated seedling counting achieved less error than the manual sampling-based practice. Regression-based counting from optical flow inputs achieved the best performance. Machine counting produces seedling density maps for management practice improvement. Abstract. The southern U.S. produces over 1 billion pine seedlings for market sales per year, with prices varying from 50 to 435 dollars per thousand seedlings. An accurate inventory of seedlings provides nursery management with insights into how many seedlings can be sold and/or if there is any loss due to washout, mechanical damage, or pest/diseases that can still be mitigated. In this study, we developed a system to count pine seedlings at production sites and map the seedling density in the field. A system with three cameras was developed to collect video from different drills in the seedling bed. The videos were preprocessed to restrict the region of interest to the center portion of the image in each camera and separated each drill into individual videos. Two different modalities, i.e., video and optical flow, were evaluated as inputs to a convolutional neural network followed by a long short-term recurrent network to model the sequence of frames and regress to the seedling count for each plot. The mean absolute percentage error (MAPE) of our best performing model was 7.5%, which is an improvement over the baseline manual sampling-based approach with a MAPE of 11%. The results showed that the proposed approach was able to count seedlings in a crowded scene under complex field conditions with higher accuracy than the standard manual practice. Therefore, the proposed system and results demonstrated the potential to replace manual counting and even provide further information such as a seedling density map over the field for precision forest nursery management and seedling harvesting. Keywords: CNN, LSTM, Nursery Inventory, Optical Flow, Pine Seedling, Regression.

British-Bangladeshi Muslim men: removing barriers to mental health support and effectively supporting our community

Abstract In England, Bangladeshi men are amongst the lowest number of people referred to primary care NHS Talking Therapies services and amongst the most likely to have deteriorated (NHS Digital, 2020). Factors related to culture, religion and gender influence stigma and help-seeking (Robinson et al., 2011). Furthermore, a lack of knowledge from services and professionals on cultural understandings of mental distress facilitate a failure to fully understand the needs of individual populations (Faheem, 2023). The aim of this research study was to qualitatively explore stigma and help-seeking for mental health difficulties, within British-Bangladeshi Muslim men in London. Recruitment consisted of men aged 22–59 years, experiencing symptoms of anxiety and/or depression but not accessing formal support. Ethical approval was obtained from Royal Holloway, University of London. After an online screening questionnaire, individual semi-structured video interviews were completed between May and October 2020. Reflexive thematic analysis presented six over-arching themes: ‘different understanding of mental health’, ‘traditional cultural expectations’, ‘fear and loss’, ‘coping resources’, ‘barriers to access’ and ‘community outreach and collaboration’. Many factors were highlighted as barriers accessing support, such as stigma and trust in services. Key practice implications are highlighted for services and practitioners. Professionals must understand the factors which impact the wellbeing of Bangladeshi men, how to better meet the needs of the community, and remove barriers to help. Participants also suggested initiatives to raise mental health awareness and reduce stigma, as an inclusive approach is needed with greater listening to communities and partnerships with local authorities. Key learning aims (1) To better understand the needs of the population and factors which impact wellbeing. (2) To consider the muti-faceted barriers to access mental health support, such as CBT, and how to address these. (3) To unpack what stigma means (internally and externally) for men in the population. (4) How to support low- and high-intensity CBT practitioners to better work therapeutically to support Bangladeshi men.

Machine Translation and Gender biases in video game localisation: a corpus-based analysis

The video game industry has been a historically gender-biased terrain due to a higher number of male protagonists and hypersexualised representations [Dietz, 1998; Downs &amp; Smith, 2010; Lynch et al., 2016]. Nowadays, echoing the debate on inclusive language, companies attempt to erase gender disparity by introducing main female characters as well as non-binary characters. From a technological point of view, even though recent studies show that Machine Translation remains largely unadopted by individual video game localisers [Rivas Ginel, 2021], multilanguage vendors are willing to invest in these tools to reduce costs [LIND, 2020]. However, the predominance of the masculine in Natural Language Processing and Machine Learning has created allocation and representation biases in Neural Machine Translation [Crawford, 2017].This paper aims to analyse the percentage of gender bias resulting from the use of Google Translate, DeepL, and SmartCat when translating in-game raw content from English into French. The games DeltaRune, The Devil's Womb and The Faces of the Forest were chosen due to the presence of non-binary characters, non-sexualized characters, and female protagonists. We compared the results in order to recount and analyse the differences between these tools' output when in terms of errors related to gender. To this end, we created a parallel corpus to compare source documents and all the translations to visualise the semantic and grammatical directions of the words embeddings [Zhou; Shi; Zhao and al, 2019] and extracted the collocations and concordance lines that represented gender identity by analysing the patterns in the source language.

An integrative review of the person‐centred and experiential therapy literature on delivering individual video counselling and psychotherapy

AbstractBased on an integrative review of the relevant literature in and on person‐centred and experiential (PCE) therapy, this study identifies and discusses the findings of 13 articles on the delivery of individual video counselling and psychotherapy. Six overarching themes were identified: the discrepancy of the experiences of clients and practitioners; the possibility of good‐quality PCE therapy (psychotherapy and/or counselling); “doing” PCE therapy online; “being” online; the role of technology; and implications for training and practice. It is suggested that all these have implications for ongoing practice and education/training, and need to be considered by practitioners, educators/trainers, and professional bodies.

Abstract 9620: External Validation of Deep Learning-Based Echocardiography Algorithms to Automate View Selection and Annotation of the Echocardiogram

Background: Manual measurements of echocardiography are time-consuming, error-prone, and highly user-dependent - all of which can be overcome with deep learning (DL) algorithms. Methods: We validated the automated DL measurement of the left ventricle (LV) and left atrial (LA) global longitudinal strain (GLS) and 20 conventional echocardiographic measurements using the Us.ai DL workflow. The workflow uses convolutional neural networks (CNNs) to automatically classify 2D videos and Doppler images with accuracies between 0.91-0.99. Separate CNNs automatically annotate and assess individual videos and images. We validated automated measurement in (1) an internal test set and (2) externally against core-lab measurements in the multinational PROMIS-HFpEF study. We compared the association of automated and manual parameters with N-terminal pro-B-type natriuretic peptide (NT-proBNP) in PROMIS-HFpEF. Results: The DL workflow successfully analysed 198 studies with an average of 7.8 minutes per study, with 100% reproducibility in the external validation set. In the internal validation set (N=86), automated LA (r=0.77, mean absolute error [MAE]=8.9%) and LV (r=0.91, MAE=2.8%) GLS showed good agreement with manual measurements. In the external dataset, automated measurements showed good agreement with the manual measurements, with an MAE range of 6.8 mL for LV volumes, 7.1% for LV ejection fraction, and 1.7 for the ratio of the mitral inflow E wave to the tissue Doppler e' wave ( Table ). Automated strain measurements showed good agreement (LV GLS: r=0.74, MAE=2.7%, and LA GLS: r=0.80, MAE=4.8%) with core lab measurements in PROMIS-HFpEF. Lastly, automated and manual measurements were similarly associated with NT-proBNP. Conclusions: DL-based algorithms can successfully classify and annotate echocardiographic studies, including LA and LV GLS. DL algorithms might increase reproducibility and democratise the use of strain measurements in clinical practice.

Use of Machine Learning to Assess Cataract Surgery Skill Level With Tool Detection

To develop a method for objective analysis of the reproducible steps in routine cataract surgery. Prospective study; machine learning. Deidentified faculty and trainee surgical videos. Consecutive cataract surgeries performed by a faculty or trainee surgeon in an ophthalmology residency program over 6 months were collected and labeled according to degrees of difficulty. An existing image classification network, ResNet 152, was fine-tuned for tool detection in cataract surgery to allow for automatic identification of each unique surgical instrument. Individual microscope video frame windows were subsequently encoded as a vector. The relation between vector encodings and perceived skill using k-fold user-out cross-validation was examined. Algorithms were evaluated using area under the receiver operating characteristic curve (AUC) and the classification accuracy. Accuracy of tool detection and skill assessment. In total, 391 consecutive cataract procedures with 209 routine cases were used. Our model achieved an AUC ranging from 0.933 to 0.998 for tool detection. For skill classification, AUC was 0.550 (95% confidence interval [CI], 0.547-0.553) with an accuracy of 54.3% (95% CI, 53.9%-54.7%) for a single snippet, AUC was 0.570 (0.565-0.575) with an accuracy of 57.8% (56.8%-58.7%) for a single surgery, and AUC was 0.692 (0.659-0.758) with an accuracy of 63.3% (56.8%-69.8%) for a single user given all their trials. Our research shows that machine learning can accurately and independently identify distinct cataract surgery tools in videos, which is crucial for comparing the use of the tool in a step. However, it is more challenging for machine learning to accurately differentiate overall and specific step skill to assess the level of training or expertise. The author(s) have no proprietary or commercial interest in any materials discussed in this article.

Multimodal recognition of frustration during game-play with deep neural networks

Frustration, which is one aspect of the field of emotional recognition, is of particular interest to the video game industry as it provides information concerning each individual player’s level of engagement. The use of non-invasive strategies to estimate this emotion is, therefore, a relevant line of research with a direct application to real-world scenarios. While several proposals regarding the performance of non-invasive frustration recognition can be found in literature, they usually rely on hand-crafted features and rarely exploit the potential inherent to the combination of different sources of information. This work, therefore, presents a new approach that automatically extracts meaningful descriptors from individual audio and video sources of information using Deep Neural Networks (DNN) in order to then combine them, with the objective of detecting frustration in Game-Play scenarios. More precisely, two fusion modalities, namely decision-level and feature-level, are presented and compared with state-of-the-art methods, along with different DNN architectures optimized for each type of data. Experiments performed with a real-world audiovisual benchmarking corpus revealed that the multimodal proposals introduced herein are more suitable than those of a unimodal nature, and that their performance also surpasses that of other state-of-the–art approaches, with error rate improvements of between 40% and 90%.

Overshadowed by shadowing: exploring how Canadian medical students experience shadowing.

Preclinical medical students commonly perceive shadowing as beneficial for career exploration. However, research is sparse on the broader impact of shadowing as a learning strategy. We explored students' perceptions and lived experiences of shadowing to understand its role and impact on their personal and professional lives. Between 2020-2021, individual semi-structured video interviews were conducted with 15 Canadian medical students in this qualitative descriptive study. Inductive analysis proceeded concurrently with data collection until no new dominant concepts were identified. Data were iteratively coded and grouped into themes. Participants described internal and external factors that moulded shadowing experiences, arising tensions between intended and perceived experiences, and how these lived experiences impacted their wellness. Internal factors associated with shadowing behaviour included: 1) aspiring to be the best and shadowing to demonstrate excellence, 2) shadowing for career exploration, 3) shadowing as learning opportunities for early clinical exposure and career preparedness, and 4) reaffirming and redefining professional identity through shadowing. External factors were: 1) unclear residency match processes which position shadowing as competitive leverage, 2) faculty messaging that perpetuates student confusion around the intended value of shadowing, and 3) social comparison in peer discourse, fuelling a competitive shadowing culture. The tension between balancing wellness with career ambitions and the unintended consequences of unclear messaging regarding shadowing in a competitive medical culture highlights issues inherent in shadowing culture.