Hours Of Footage Research Articles

HabitSense: A Privacy-Aware, AI-Enhanced Multimodal Wearable Platform for mHealth Applications

Wearable cameras provide an objective method to visually confirm and automate the detection of health-risk behaviors such as smoking and overeating, which is critical for developing and testing adaptive treatment interventions. Despite the potential of wearable camera systems, adoption is hindered by inadequate clinician input in the design, user privacy concerns, and user burden. To address these barriers, we introduced HabitSense, an open-source1, multi-modal neck-worn platform developed with input from focus groups with clinicians (N=36) and user feedback from in-wild studies involving 105 participants over 35 days. Optimized for monitoring health-risk behaviors, the platform utilizes RGB, thermal, and inertial measurement unit sensors to detect eating and smoking events in real time. In a 7-day study involving 15 participants, HabitSense recorded 768 hours of footage, capturing 420.91 minutes of hand-to-mouth gestures associated with eating and smoking data crucial for training machine learning models, achieving a 92% F1-score in gesture recognition. To address privacy concerns, the platform records only during likely health-risk behavior events using SECURE, a smart activation algorithm. Additionally, HabitSense employs on-device obfuscation algorithms that selectively obfuscate the background during recording, maintaining individual privacy while leaving gestures related to health-risk behaviors unobfuscated. Our implementation of SECURE has resulted in a 48% reduction in storage needs and a 30% increase in battery life. This paper highlights the critical roles of clinician feedback, extensive field testing, and privacy-enhancing algorithms in developing an unobtrusive, lightweight, and reproducible wearable system that is both feasible and acceptable for monitoring health-risk behaviors in real-world settings.

Chelonia mydas detection and image extraction from noisy field recordings

<span lang="EN-US">Wildlife videography is an essential data collection method for conducting research on animals. The video recording process of an animal like the <em>Chelonia Mydas</em> turtle in its natural habitat requires the setting up of special camera traps or by performing complex camera movement to capture the animal in frame whilst the cameraman maneuvers over uneven terrain while filming. The result is hours of footage that only have the presence of the intended subject in it for seconds whilst the rest is background footage; or noisy and blurry footage that has only several usable frames among thousands of noisy and unusable ones. This presents a problem that deep learning models can help to assist, especially in detecting a wildlife subject and extracting usable data from hours of noise and background footage. This paper proposes the use of machine learning models to detect and extract wildlife images of <em>Chelonia Mydas</em> turtles to help prune through hundreds and thousands of frames from several video footages. Our paper shows that utilizing a custom model with various confidence scores can label and crop out images in noisy field video recordings of </span><em><span lang="EN-US">Chelonia Mydas </span></em><span lang="EN-US">turtles with up to 99.89% of output images correctly cropped and labeled.</span>

DeepSum: A Deep Learning Framework for Summarizing Animal Behavior

The burgeoning field of ethology necessitates efficient tools for analyzing extensive video recordings of animal behavior, as manually sifting through hours of footage is both time-consuming and susceptible to observer bias. Here we present an innovative deep learning framework tailored for summarizing animal behavior videos, aiming to distill lengthy recordings into concise, informative segments. Leveraging the latest advancements in hierarchical video summarization, our approach employs a combination of Convolutional Neural Networks (CNNs) and Transformer models to extract and understand complex spatial-temporal patterns inherent in animal movements and interactions. The model is designed to recognize and prioritize key behavioral events, ensuring the retention of critical moments in the summarized output. Additionally, an attention mechanism is incorporated to adaptively focus on salient features, enhancing the model’s capability to discern subtle yet significant behavioral nuances. We assess our framework on a range of datasets containing different species and behavioral situations, and find that it outperforms current state-of-the-art techniques in terms of accuracy, coherence, and informativeness of the generated summaries. In addition to providing a consistent, objective method of analyzing animal behavior, DeepSum dramatically reduces the amount of manual labor needed for behavioral analysis, opening the door for advancements in ethological research and wildlife conservation.

Context-Driven Detection of Invertebrate Species in Deep-Sea Video

Each year, underwater remotely operated vehicles (ROVs) collect thousands of hours of video of unexplored ocean habitats revealing a plethora of information regarding biodiversity on Earth. However, fully utilizing this information remains a challenge as proper annotations and analysis require trained scientists’ time, which is both limited and costly. To this end, we present a Dataset for Underwater Substrate and Invertebrate Analysis (DUSIA), a benchmark suite and growing large-scale dataset to train, validate, and test methods for temporally localizing four underwater substrates as well as temporally and spatially localizing 59 underwater invertebrate species. DUSIA currently includes over ten hours of footage across 25 videos captured in 1080p at 30 fps by an ROV following pre-planned transects across the ocean floor near the Channel Islands of California. Each video includes annotations indicating the start and end times of substrates across the video in addition to counts of species of interest. Some frames are annotated with precise bounding box locations for invertebrate species of interest, as seen in Fig. 1. To our knowledge, DUSIA is the first dataset of its kind for deep sea exploration, with video from a moving camera, that includes substrate annotations and invertebrate species that are present at significant depths where sunlight does not penetrate. Additionally, we present the novel context-driven object detector (CDD) where we use explicit substrate classification to influence an object detection network to simultaneously predict a substrate and species class influenced by that substrate. We also present a method for improving training on partially annotated bounding box frames. Finally, we offer a baseline method for automating the counting of invertebrate species of interest.

Bud Powell Behind the Scenes

The Francis Paudras collection on Bud Powell contains over twenty hours of footage of the renowned American jazz pianist Bud Powell (1924-1966). This article explores the major highlights from the Francis Paudras collection and outlines how archivists at the Institute are making the films more accessible to researchers.

Personalized Energy Expenditure Estimation: Visual Sensing Approach With Deep Learning.

BackgroundCalorimetry is both expensive and obtrusive but provides the only way to accurately measure energy expenditure in daily living activities of any specific person, as different people can use different amounts of energy despite performing the same actions in the same manner. Deep learning video analysis techniques have traditionally required a lot of data to train; however, recent advances in few-shot learning, where only a few training examples are necessary, have made developing personalized models without a calorimeter a possibility.ObjectiveThe primary aim of this study is to determine which activities are most well suited to calibrate a vision-based personalized deep learning calorie estimation system for daily living activities.MethodsThe SPHERE (Sensor Platform for Healthcare in a Residential Environment) Calorie data set is used, which features 10 participants performing 11 daily living activities totaling 4.5 hours of footage. Calorimeter and video data are available for all recordings. A deep learning method is used to regress calorie predictions from video.ResultsModels are personalized with 32 seconds from all 11 actions in the data set, and mean square error (MSE) is taken against a calorimeter ground truth. The best single action for calibration is wipe (1.40 MSE). The best pair of actions are sweep and sit (1.09 MSE). This compares favorably to using a whole 30-minute sequence containing 11 actions to calibrate (1.06 MSE).ConclusionsA vision-based deep learning energy expenditure estimation system for a wide range of daily living activities can be calibrated to a specific person with footage and calorimeter data from 32 seconds of sweeping and 32 seconds of sitting.

“There's really almost nobody there who's trying to make sense of it”

“There's really almost nobody there who's trying to make sense of it”

The Mountain Electric

The Mountain Electric is an audiovisual ethnography of the Mountain Skies music festival held in Black Mountain, North Carolina (USA) on May 3-5, 2019. This film explores experimental ambient electronic music as a community of practice connected across virtual and physical worlds. It draws on over 30 ethnographic interviews with artists and more than 80 hours of footage.

Visual Storytelling for Knowledge Translation: A Study on BRAC’s Novel Health Loans in Protecting the Poor Against Health and Asset Vulnerability In Bangladesh

Knowledge Translation (KT) is a dynamic and iterative process that includes synthesizing, disseminating, exchanging, and ethically sound application of knowledge to improve health and strengthen the health care system. It facilitates sharing the information generated through research outcomes with the public, the policymakers, or others for further scaling up or continuation of the interventions. Literature suggests a substantial gap exists in communicating with the decision-makers. BRAC JPGSPH produced a documentary/video that iterates how BRAC’s revised medical treatment loan program (MTL+) works with its microcredit clients and modalities. After reviewing all possible options for communication, the video was chosen as the best knowledge translation tool. The video creation and dissemination process are comprised of four phases: pre-production, production, post-production, and exhibition. The video production team reviewed documents and articles and conducted multiple interviews before developing the script. Later, a series of interviews were taken with the beneficiaries who receive medical treatment loans, mid-level, senior managers at BRAC, and researchers. After the production, the director, with a professional editor, edited the video. Over three hours of footage was viewed and ultimately compiled into a six-minute-long video documentary. The audience for the video was more expansive than narrow; from potential beneficiaries to policymakers and every group of stakeholders in between, the video was well-understood. The new MTL+ was integrated into the main program and would be scaled up soon.

Study of automated shuttle interactions in city traffic using surrogate measures of safety

Driving automation is happening at a rapid pace, with different driver assistance systems already available in mass-market cars. However, this rapid development in driving automation leads to concerns and questions about their impact on safety. While previous studies have been restricted to incident reports and simulation tools, the safety of automated vehicles (AVs) has not been clearly demonstrated independently from their developers and proponents. Instead of crashes, which are extremely rare events, this study uses surrogate measures of safety (SMoS) to analyze the interactions between road users and low-speed automated shuttles that circulated in Montreal and Candiac, in Canada, during two pilot projects in mid and late 2019. Cameras were placed at seven intersections along the routes of the shuttles. More than 70 hours of footage were processed to extract the road user trajectories using computer vision techniques and compute various safety indicators: speed and acceleration at the road user level, time headway, time-to-collision (TTC), post-encroachment time (PET) and the speed difference at the interaction level. The Kolmogorov–Smirnov test was used to compare the distributions of these indicators for interactions involving automated shuttles or motorized vehicles following similar paths. Multivariate regression was used to identify the relationship between the safety indicators and various factors. The results indicate that these automated shuttles behave generally more safely: their speeds and accelerations are lower and their interactions are characterized by higher TTC and PET values, and lower speed differences. However, small headway times at one site with high-speed differences between the shuttles and other following vehicles raise concerns that warrant further research into the suitable context for these vehicles.

Can independent regulatory agencies mend Europe’s democracy? The case of the European Medicines Agency’s public hearing on Valproate

In 2017, the European Medicines Agency staged the first effort at democratic innovation within transnational European Union institutions directly influencing the transnational regulation of medicines. Alongside its public consultation on epilepsy drug Valproate, European Medicines Agency included a public hearing involving representatives of patients and testimony from those directly affected by the medicines. Using this critical case study, the article argues from a deliberative democratic perspective that although the hearing in many ways exhibited the traditional shortcomings of elite-driven deliberation, it also demonstrated unexpected and surprising deliberative qualities. Based on new quantitative analysis of the hearing using the Discourse Quality Index, and qualitative observation of over 4 hours of footage, the article argues European Medicines Agency’s hearing facilitated equitable access and influence of patients and members of the public who had previously been excluded from decision making on drug distribution. These findings provide important new evidence of the deliberative democratic value of public hearings.

Torn limb from limb: the ethology of prey-processing in Tasmanian devils (

The success of carnivorous mammals is determined not only by their ability to locate and kill prey, but also their efficiency at consuming it. Breaking large prey into small pieces is challenging due to the strong and tough materials that make up a carcass (e.g. hide, muscle, and bone). Carnivores therefore require a diverse suite of prey-processing behaviours to utilise their catch. Tasmanian devils are Australia’s only large marsupial scavengers and have the ability to consume almost all of a carcass. To determine how they do this we analysed 5.5 hours of footage from 21 captive and wild devils feeding at carcasses. We documented 6320 bouts of 12 distinct prey-processing behaviours, performed at frequencies that varied throughout feeds and between groups. The time point in the feed influenced the types of behaviours used. This is likely due to changing prey size, as different techniques appear better suited to handling whole carcasses or large pieces (pulling and pinning) or smaller pieces (holding and manipulating). Group size impacted the frequency of social pulling behaviours, which increased with the number of animals. Our findings highlight the range of prey-processing behaviours performed by scavenging devils when handling, breaking down, and consuming a carcass. The devils’ repertoire shares similarities with large carnivores that handle and consume whole carcasses as well as small carnivores that are adept in grasping and handling smaller prey.

For Sama

For Sama

Double Stimulation and Transformative Agency through Video Modalities with Young Children: A cross-national study of pedagogic relationships in families

ABSTRACT Our research community has limited understanding about the potential of video. This paper uses double stimulation and transformative agency to analyse two video modalities, extraction, and reflection; in a study about learning as it develops through family interactions. As researchers we were interested in children’s motives. However, the children were young; therefore, posing a methodological problem. Drawing on video footage and 24 film elicitation interviews with children and fathers from 12 families in Norway, England, Hong Kong and Mumbai, India; we present case study examples using double stimulation to analyse video. We gathered between 5-10 hours of footage about father-child interactions from each family. To analyse video modalities through double stimulation, we identified the first stimulus, second stimulus, and manifestations of transformative agency. We found by combining double stimulation with video for the purposes of extraction, the non-visible becomes visible. The combination indicates opportunities for development that generate the kind of qualitative transformations, which imply child development can be seen in a new way. By combining double stimulation with video for reflection, researchers can trace outward affective and cognitive consequences of the ‘mirror effect’, back to children’s conflict of motives, and how they decide to interpret what they see. Double stimulation therefore widens the potential of video. It shows video as an effective auxiliary stimulus to address the problem of researching young children’s motives.

A Poetic of Beauty: Nature, Memory and Resilience in El Botón de Nacar/The Pearl Button (2015)

One of the pillars of the history of Latin American documentary filmmaking is undoubtedly the Chilean director Patricio Guzmán. During more than five decades of film production, he has given an account of Chile's political history, specifically dealing with the meaning of the Chilean dictatorship for the country's past, present and future, as well as its relevance in the Latin American context. His films undoubtedly represent one of the most relevant testimonies to the continent's history in the 20th century, with hundreds of hours of footage of the high points in the violent transition process between the Allende and Pinochet governments. His film proposal will take a radical turn in his last two productions: Nostalgia de la Luz ([2010] 2011) and El Botón de Nácar (2015), which are part of a trilogy - the latter film is currently in production. Its aesthetic proposal is revolutionary in the sense that it resignifies the relationship between memory and history by establishing a connection between nature, cosmology and historical memory. In this trilogy he deals separately in each film with the desert, the sea and the mountain, leading the viewer to a reflection that, through poetic language and a meditative tone, proposes a new look at history and memory. In this essay I analyze three elements that define his film El Botón de Nacar and that open the possibilities for a new type of theoretical reflection on the documentary genre: the relationship between nature, narration and history, history understood as genealogical memory and the new type of visual aesthetic that the director is proposing, which I call "poetry of beauty".

The Edge of Democracy

Reviewed by: The Edge of Democracy (Democracia em Vertigem) by Petra Costa Sofia Paiva de Araujo (bio) The Edge of Democracy (Democracia em Vertigem). Petra Costa, 2019. Netflix. 123 mins. Petra Costa’s documentary The Edge of Democracy presents the 2016 impeachment of Brazilian president Dilma Rousseff as a continuation of the old politics that led to the civil-military coup of 1964. It interprets the impeachment as a result of the institutional amnesia that followed the democratic transition initiated in 1979. The lack of punishment of crimes committed by the dictatorship is directly related to the deeply ingrained promiscuous relationship between the nation’s political and economic elites. Costa links the nonpunishment of the human rights abuses of the dictatorial period to the rise of the far-right president Jair Bolsonaro. The documentary also relates the latter to the growing influence of evangelical fundamentalists and armed militant groups in Brazilian politics. Nevertheless, the uniqueness of the documentary lies in Costa’s approach to address such sensitive and complex issues. Using the formula of a documentary essay, Costa intentionally blurs the line between memoir and journalism, inserting herself in the narrative. The Edge of Democracy is personal, original, and passionate. Costa brings a sui generis perspective to the table by taking advantage of the complex ambiguities of her family’s background, which she ingeniously inter-weaves with major events of Brazilian history in a heartfelt first-person voice-over. On one hand, her maternal family is a token for the privileged economic elites who supported the 1964 coup and have since then shaped politics by funding politicians who enact their agenda and safeguard their interests. On the other hand, her parents’ political activism during the dictatorship led to their arrest and years of living in hiding, making them a token for the idealistic youth who put themselves at risk by fighting against an authoritarian and violent regime. Furthermore, Costa adds another layer of reflection by making an analogy between the development of Brazilian politics in the post—dictatorship era and her own process of growing political awareness: the director and Brazilian democracy are roughly the same age and have developed and matured together. Costa grew up believing Brazilian democracy was a right earned through the political activism of her parents’ generation. As she grew older and more critical, the director realized this right could not be taken for granted, as Brazil’s young democracy is fragile and vulnerable to power games. [End Page 187] The originality of the documentary results from a thorough interpretation of Brazilian politics crafted from a careful collage of more than 6,000 hours of footage. While some of the scenes use mainstream media news, most of the film shows footage collected by the director and her team, who made great use of the privileged social background and connections of Costa and her family. They walked the hallways of the House of Representatives, the Congress, and the president’s official residence; they interviewed politicians from both sides of the spectrum and closely followed Dilma Rousseff and her predecessor, Luis Inácio Lula da Silva, before, during, and after the impeachment. As a result, The Edge of Democracy is a firsthand account of someone who actively witnessed “behind the scenes” negotiations in Brasília and passionately examined the origins of the country’s political crisis by quilting together the pieces of evidence she had gathered. However, Costa’s closeness to the leaders of the Workers’ Party led to criticism from both the Left and the Right, and the documentary received mixed reviews in Brazil. On one hand, people leaning Right on the political spectrum perceived The Edge of Democracy as biased—especially Bolsonaro’s supporters, who saw it as straightforward propaganda. On the other hand, members of the far Left used Costa’s family’s connections with the infamous construction company Andrade-Gutierrez to discredit the value of her argument—despite the fact that Costa is aware of her privileges and openly critical of the company’s involvement in politics and corruption scandals. Domestically, the documentary’s positive reviews came mainly from Workers’ Party militants and people in the center-Left and liberal Left. They praised Costa...

Viviane Reding on her action in the field of the information society and media (2004-2010) Interview by Elena Danescu

The following pages are taken from a long interview (more than eight hours of footage in total) that Viviane Reding granted us in 2015 in connection with the “Pierre Werner and Europe” research project. Drawing on more than 40 years’ experience in politics, Viviane Reding spoke about her career, the role of Luxembourg and Luxembourgers in the European integration process, and various key events in which she played a part. In these extracts, she discusses her role as a member of the first and second Barroso Commissions (2004-2009 and 2010-2014) and her efforts to help build an information and knowledge society in Europe, one that serves citizens and protects their rights and fundamental freedoms. Her achievements in this respect include capping mobile phone roaming charges (they were subsequently abolished in 2017), advocating for the introduction of a single emergency number (112) in all EU countries, launching the Europeana digital library, and spearheading a programme to use technological innovation for climate and energy solutions. She also describes the process of developing a Digital Agenda for Europe to improve the continent’s digital competitiveness compared with the United States, China and Japan – a complex and challenging task given the context of globalisation and the divergent interests of the various stakeholders (research, industry, consumers, etc.). Finally, she mentions the reform of personal data protection that she initiated (leading to the GDPR, adopted in April 2016).

Activity and social interactions in a wide-ranging specialist scavenger, the Tasmanian devil (Sarcophilus harrisii), revealed by animal-borne video collars

Observing animals directly in the field provides the most accurate understanding of animal behaviour and resource selection. However, making prolonged observation of undisturbed animals is difficult or impossible for many species. To overcome this problem for the Tasmanian devil (Sarcophilus harrisii), a cryptic and nocturnal carnivore, we developed animal-borne video collars to investigate activity patterns, foraging behaviour and social interactions. We collected 173 hours of footage from 13 individual devils between 2013 and 2017. Devils were active mostly at night, and resting was the most common behaviour in all diel periods. Devils spent more time scavenging than hunting and exhibited opportunistic and flexible foraging behaviours. Scavenging occurred mostly in natural vegetation but also in anthropogenic vegetation and linear features (roads and fence lines). Scavenging frequency was inversely incremental with size e.g. small carcasses were scavenged most frequently. Agonistic interactions with conspecifics occurred most often when devils were traveling but also occurred over carcasses or dens. Interactions generally involved vocalisations and brief chases without physical contact. Our results highlight the importance of devils as a scavenger in the Tasmanian ecosystem, not just of large carcasses for which devils are well known but in cleaning up small items of carrion in the bush. Our results also show the complex nature of intraspecific interactions, revealing greater detail on the context in which interactions occur. In addition, this study demonstrates the benefits of using animal-borne imaging in quantifying behaviour of elusive, nocturnal carnivores not previously seen using conventional field methods.

Mapping subtidal estuarine habitats with a remotely operated underwater vehicle (ROV)

Subtidal habitats have not yet been accounted for in habitat maps of South African estuaries. In this study, a novel method for mapping subtidal estuarine habitats, using a remotely operated underwater vehicle (ROV) piloted from a boat, was developed and tested in the Knysna Estuary. Video footage was recorded along 48 transects across the width of the estuary, and then reviewed to identify, classify and map habitats. Using the method developed in this study, 21 hours of footage was recorded over 15 days of sampling, and about 30 hours of post-processing was carried out to map an area exceeding 850 ha. This study has produced the first baseline dataset of subtidal habitats for a South African estuary. Additionally, the study revealed the previously unknown distribution of the invasive red seaweed Asparagopsis taxiformis, and the underestimation in previous studies of the estuary of area cover of eelgrass Zostera capensis by 130 ha.

Key wording practices in three aphasia conversation groups: a preliminary study

ABSTRACT Background: When facilitating an aphasia group conversation, many facilitators engage in key wording or the action of writing down brief notes on the conversation. Though available sources provide general information about key wording and its benefits, no observational, interaction-oriented studies of key wording currently exist. Cognitive ethnography (CE), a trans-disciplinary field which examines how human beings create and engage in complex activity, was selected as the guiding paradigm for this study. Aim: The aim of this preliminary, qualitative study was to examine and document the process of key wording as it occurs in three aphasia conversation groups. Method and Procedure: We reviewed fifteen hours of footage of three aphasia conversation groups (Sites A, B, and C) several times, focusing on segments of interaction in which key wording took place. Tanith (Site A), Leo (Site B), and Adelaide (Site C) were the facilitators we observed. As we noted patterns of behavior in the groups’ use of key wording, we wrote a series of memoranda to record our developing understandings of the key wording process. We completed a cycle of writing and comparing the ideas expressed in the memoranda against the data numerous times. We employed several verification procedures, namely, triangulation, member checking, and lamination. Outcomes and Results: All three groups engaged in key wording many times during each session. The groups we observed used systems composed of different elements to engage in key wording. One group relied on distribution of tasks across participants in that the facilitator attended to other tasks related to managing the interaction while a co-facilitator provided individual key wording support to group members with auditory comprehension deficits. In another group, the facilitator produced a log of the conversation on a sequence of pages located in the middle of the table, where all group members could see them. Finally, the facilitators and members of the third group collaboratively created graphic organizers on a white board which was located in the front of the room. Conclusion: In the three groups we observed in this study, key wording was a routine process which involved the distribution of tasks across participants and the use of external devices. The available interactional data suggest that the inscriptions facilitators produce help group members with aphasia comprehend various local and global aspects of ongoing conversation.