Video Analysis System Research Articles

P0311 Artificial intelligence methods for informative frame detection reduce central reading time and improve disease grading in Ulcerative Colitis clinical trials

Abstract Background Central reading using endoscopic scoring systems is recommended in clinical trials to assess disease severity in ulcerative colitis (UC). Human scoring of the Mayo Endoscopic Subscore (MES) of colonoscopy videos is time-consuming and can have a high inter-reader variability. The automatic extraction of salient frames from endoscopic videos can assist the development of computer-aided scoring, reduce local and central reader discrepancies and allow for a faster review process [1, 2]. We present an informative frame selection framework prioritizing the video segments most representative of disease activity, accelerating manual endoscopic scoring and enhancing video analysis systems. Methods An artificial intelligence (AI) system based on neural networks was trained on 104,572 endoscopic frames to automatically detect four image artifacts (blurriness, out-of-body, camera movements and obstructed views). A new temporal smoothing module enables full-length video analysis leveraging still-frame classification confidence levels. The system was evaluated on an independent dataset of retrospectively acquired videos from a large multicentric cohort (1,564 UC subjects, total video time: 357 hours). Videos were annotated for MES grading by six expert gastroenterologists using a 2 + 1 paradigm. Mean non-informative time and mean video informativeness level (mVIL) were evaluated to estimate the projected time savings by filtering non-representative frames. The system was integrated in a previously presented AI-powered reading platform, enabling video editing by removal of non-assessable intervals. Quadratic weighted kappa (QWK) was evaluated to determine the effect of video trimming on MES scoring by comparing the AI platform accuracy to manual annotations. Results Accuracy in detecting non-informative frames was 80.02% based on 12,575 independent samples. Non-informative filtering allowed for a mean reduction in video time of 3:09 minutes, corresponding to a mVIL of 77.61% and a total timesaving of 83 hours. When used in conjunction with an automatic MES grading system, very high correlation was observed between gastroenterologists and predicted scores both when evaluated over the whole dataset (QWK=0.690) and for concordant central and local readings (QWK=0.725). Video filtering improved the automatic MES classification accuracy by 1.1%. Conclusion The presented automatic AI system significantly reduces endoscopic video length by filtering non-informative frames and improves performance of automated MES grading algorithms. When tested on a large cohort of UC subjects, the system demonstrated the potential to reduce manual scoring time and reader fatigue without compromising objective MES assessment.

A deep learning based detection algorithm for anomalous behavior and anomalous item on buses

This paper proposes a new strategy for analysing and detecting abnormal passenger behavior and abnormal objects on buses. First, a library of abnormal passenger behaviors and objects on buses is established. Then, a new mask detection and abnormal object detection and analysis (MD-AODA) algorithm is proposed. The algorithm is based on the deep learning YOLOv5 (You Only Look Once) algorithm with improvements. For onboard face mask detection, a strategy based on the combination of onboard face detection and target tracking is used. To detect abnormal objects in the vehicle, a geometric scale conversion-based approach for recognizing large-size ab-normal objects is adopted. To apply the algorithm effectively to real bus data, an embedded video analysis system is designed. The system incorporates the proposed method, which results in improved accuracy and timeliness in detecting anomalies compared to existing approaches. The algorithm’s effectiveness and applicability is verified through comprehensive experiments using actual video bus data. The experimental results affirm the validity and practicality of the pro-posed algorithm.

IMU Sensor-Based Worker Behavior Recognition and Construction of a Cyber-Physical System Environment.

According to South Korea's Ministry of Employment and Labor, approximately 25,000 construction workers suffered from various injuries between 2015 and 2019. Additionally, about 500 fatalities occur annually, and multiple studies are being conducted to prevent these accidents and quickly identify their occurrence to secure the golden time for the injured. Recently, AI-based video analysis systems for detecting safety accidents have been introduced. However, these systems are limited to areas where CCTV is installed, and in locations like construction sites, numerous blind spots exist due to the limitations of CCTV coverage. To address this issue, there is active research on the use of MEMS (micro-electromechanical systems) sensors to detect abnormal conditions in workers. In particular, methods such as using accelerometers and gyroscopes within MEMS sensors to acquire data based on workers' angles, utilizing three-axis accelerometers and barometric pressure sensors to improve the accuracy of fall detection systems, and measuring the wearer's gait using the x-, y-, and z-axis data from accelerometers and gyroscopes are being studied. However, most methods involve use of MEMS sensors embedded in smartphones, typically attaching the sensors to one or two specific body parts. Therefore, in this study, we developed a novel miniaturized IMU (inertial measurement unit) sensor that can be simultaneously attached to multiple body parts of construction workers (head, body, hands, and legs). The sensor integrates accelerometers, gyroscopes, and barometric pressure sensors to measure various worker movements in real time (e.g., walking, jumping, standing, and working at heights). Additionally, incorporating PPG (photoplethysmography), body temperature, and acoustic sensors, enables the comprehensive observation of both physiological signals and environmental changes. The collected sensor data are preprocessed using Kalman and extended Kalman filters, among others, and an algorithm was proposed to evaluate workers' safety status and update health-related data in real time. Experimental results demonstrated that the proposed IMU sensor can classify work activities with over 90% accuracy even at a low sampling rate of 15 Hz. Furthermore, by integrating internal filtering, communication modules, and server connectivity within an application, we established a cyber-physical system (CPS), enabling real-time monitoring and immediate alert transmission to safety managers. Through this approach, we verified improved performance in terms of miniaturization, measurement accuracy, and server integration compared to existing commercial sensors.

In-vehicle edge system for real-time dashcam video analysis

In-vehicle edge system for real-time dashcam video analysis

출입자 안전장구 검출 시스템 연구

The safety of workers at industrial sites is a critical issue, and various facilities and regulations are in place to ensure safety. Among these, the most important factor is raising field workers' awareness of wearing safety equipment. However, many workers neglect to wear such equipment due to inconvenience or the demands of their tasks. Furthermore, with the expansion of the *Serious Accident Punishment Act* to include workplaces with fewer than 50 employees starting in 2024, the establishment and implementation of a safety and health management system led by management have become essential.In response to this, this paper proposes a system that can detect whether workers are wearing safety gear upon entering the worksite. It also explores the development of an operational environment to encourage compliance and validate its effectiveness through field testing. Building a video analysis system powered by a high-end GPU, however, can impose a significant financial burden. To address this, the study introduces a solution leveraging an edge device-based system for intelligent video analysis, aiming to create a product that can be easily deployed in the field.

High Agreement Between 3D Motion Capture and 2D Video Analysis Systems and Raters for Knee Flexion Range of Motion During a Single-Leg Drop Landing Task

OBJECTIVE: To determine the agreement between a two-dimensional (2D) and three-dimensional (3D) video capture system and between multiple raters during a single-leg drop landing task. DESIGN: Cross-sectional observational study. METHODS: Single-leg drop landing trials of 23 males (n = 12) and females (n = 11) were recorded using the 2D and 3D systems. Reliability and agreement were quantified through intraclass correlation coefficients (ICC2, k), standard error of the measurement (SEM), and Bland-Altman plots. Analysis of variance assessed systematic error between the 2D raters and 3D system. RESULTS: A 1-way analysis of variance on 35 drop-landing trials across raters detected a difference across the 3 estimates of knee flexion range of motion ( P<.001) where rater 1 did not differ (mean difference ± standard error: 1.15° ± 0.48°, P = .54) from the 3D system, rater 2 > 3D system (3.30° ± 0.76°, P<.001), and raters 1 and 2 differed (2.14° ± 0.82°, P = .04). There was high reliability between the two 2D raters (ICC2, k = .89, SEM = 3.44°). Rater 1 completed analyses on 87 additional trials (total n = 117) and was comparable to the 3D system (0.87° ± 0.47°, P = .06; ICC2, k = .90; SEM = 3.52°). CONCLUSION: Two-dimensional video analysis can be used in a clinical setting where 3D capture of movement mechanics is not feasible. JOSPT Open 2025;3(1):1-8. Epub 10 September 2024. doi:10.2519/josptopen.2024.0051

A Coding Framework and Benchmark Towards Low-Bitrate Video Understanding.

Video compression is indispensable to most video analysis systems. Despite saving the transportation bandwidth, it also deteriorates downstream video understanding tasks, especially at low-bitrate settings. To systematically investigate this problem, we first thoroughly review the previous methods, revealing that three principles, i.e., task-decoupled, label-free, and data-emerged semantic prior, are critical to a machine-friendly coding framework but are not fully satisfied so far. In this paper, we propose a traditional-neural mixed coding framework that simultaneously fulfills all these principles, by taking advantage of both traditional codecs and neural networks (NNs). On one hand, the traditional codecs can efficiently encode the pixel signal of videos but may distort the semantic information. On the other hand, highly non-linear NNs are proficient in condensing video semantics into a compact representation. The framework is optimized by ensuring that a transportation-efficient semantic representation of the video is preserved w.r.t. the coding procedure, which is spontaneously learned from unlabeled data in a self-supervised manner. The videos collaboratively decoded from two streams (codec and NN) are of rich semantics, as well as visually photo-realistic, empirically boosting several mainstream downstream video analysis task performances without any post-adaptation procedure. Furthermore, by introducing the attention mechanism and adaptive modeling scheme, the video semantic modeling ability of our approach is further enhanced. Fianlly, we build a low-bitrate video understanding benchmark with three downstream tasks on eight datasets, demonstrating the notable superiority of our approach. All codes, data, and models will be open-sourced for facilitating future research.

Influence of Substitutions and Roles on Kinematic Variables in Professional Soccer Players.

In soccer, roles and substitutions can make it unclear how different physical capacities decrease over time and impact performance. This investigation aimed to analyze kinematic parameters and their changes over game time, and provide information to effectively prescribe training programs. Data from four professional teams were analyzed, and all the teams competed in the Italian First Division (Serie A). Average metabolic power and high-intensity activities associated with critical moments in the match were considered in the comparisons. The video analysis system STATS SportVU collected the data during official matches. Players were assigned to a specific group according to their time played, categorized as follows: 0-15', 15-30', 30-45', 45-60', 60-75', 75-90', and > 90 min. Different roles, including forwards and wingers (FW), midfielders (MF), center-backs (CB), and side-backs (SB) were also considered. Significant differences (p<0.05) were found in different roles and within roles at different times played. The results highlighted how MF performance decreased over time, whereas CB was unaffected by time. SB spent more time at very high velocity than other roles, independently from the time played. These findings provide valuable information about the physical demands of official soccer matches and could be used to review training prescriptions.

Performance in professional soccer players: normative data and benchmarks from official matches for metabolic power and high-intensity activities.

The opportunity to compare data among players using normative data and benchmarks may represent a helpful tool for coaches to assess the players' physical performance easily and quickly. This study aimed to create normative data and benchmarks that can be exploited to compare professional soccer players' physical capacity competing in the Premier Division Championship (series A). Match data from six professional soccer teams competing in the Premier Division championship (Serie A and Italy Cup matches) were collected during the entire season. Players (N.=112) were divided based on the role positioning as follows: forwards and wingers (FW), midfielders (MF), side-backs (SB), and center-backs (CB). All the teams analyzed competed with a 4-3-3 formation, and only players who played for the entire match (85-95 minutes) were considered. The video analysis system STATS SportVU was used to collect the data during official matches. Average metabolic power (AP) was considered to estimate the energy cost. The number of power events (PE), distance (m) covered at more than 25 W/kg (D25), and finally, the distance covered at v>25 km/h (VHS) were considered as high-intensity assessments. Standardized T-scores (scaled from 0 to 100) were calculated for each role and variable. T-score data was used to create performance bands combined with qualitative description (ranging from extremely poor to excellent), and a traffic light system approach was implemented to simplify the data's interpretation. The results could be used to compare different athletes' performance quickly and effectively, to detect symptoms of overtraining, and to give helpful insights to coaches on what the training should be focused on.

Intelligent Video Monitoring and Analysis System for Power Grid Construction Site Safety Using Wireless Power Transfer

Power grid construction significantly enhances power grid management and risk control. Inconsistent operations at construction sites can jeopardize grid stability and crew safety. Traditional power lines are less favored due to mobility limitations, while batteries add burden and impracticality. To address this, a wireless power transfer and video monitoring system is developed using RF technology and Yolo V3 model. This enables continuous monitoring and employee safety analysis. The system's detection performance is optimized using HPSO, surpassing existing methods in accuracy and speed. It ensures real-time monitoring and improves the detection of potential risk sources, crucial for construction site safety.

Collaborative Human-Computer Vision Operative Video Analysis Algorithm for Analyzing Surgical Fluency and Surgical Interruptions in Endonasal Endoscopic Pituitary Surgery: Cohort Study.

The endonasal endoscopic approach (EEA) is effective for pituitary adenoma resection. However, manual review of operative videos is time-consuming. The application of a computer vision (CV) algorithm could potentially reduce the time required for operative video review and facilitate the training of surgeons to overcome the learning curve of EEA. This study aimed to evaluate the performance of a CV-based video analysis system, based on OpenCV algorithm, to detect surgical interruptions and analyze surgical fluency in EEA. The accuracy of the CV-based video analysis was investigated, and the time required for operative video review using CV-based analysis was compared to that of manual review. The dominant color of each frame in the EEA video was determined using OpenCV. We developed an algorithm to identify events of surgical interruption if the alterations in the dominant color pixels reached certain thresholds. The thresholds were determined by training the current algorithm using EEA videos. The accuracy of the CV analysis was determined by manual review, and the time spent was reported. A total of 46 EEA operative videos were analyzed, with 93.6%, 95.1%, and 93.3% accuracies in the training, test 1, and test 2 data sets, respectively. Compared with manual review, CV-based analysis reduced the time required for operative video review by 86% (manual review: 166.8 and CV analysis: 22.6 minutes; P<.001). The application of a human-computer collaborative strategy increased the overall accuracy to 98.5%, with a 74% reduction in the review time (manual review: 166.8 and human-CV collaboration: 43.4 minutes; P<.001). Analysis of the different surgical phases showed that the sellar phase had the lowest frequency (nasal phase: 14.9, sphenoidal phase: 15.9, and sellar phase: 4.9 interruptions/10 minutes; P<.001) and duration (nasal phase: 67.4, sphenoidal phase: 77.9, and sellar phase: 31.1 seconds/10 minutes; P<.001) of surgical interruptions. A comparison of the early and late EEA videos showed that increased surgical experience was associated with a decreased number (early: 4.9 and late: 2.9 interruptions/10 minutes; P=.03) and duration (early: 41.1 and late: 19.8 seconds/10 minutes; P=.02) of surgical interruptions during the sellar phase. CV-based analysis had a 93% to 98% accuracy in detecting the number, frequency, and duration of surgical interruptions occurring during EEA. Moreover, CV-based analysis reduced the time required to analyze the surgical fluency in EEA videos compared to manual review. The application of CV can facilitate the training of surgeons to overcome the learning curve of endoscopic skull base surgery. ClinicalTrials.gov NCT06156020; https://clinicaltrials.gov/study/NCT06156020.

Study on Integrated Systems for Enhanced Forest Fire Prevention: An Embedded Intelligent Video Analysis and Collaborative IoT–Fog–Cloud Framework Approach

Introduction:: Forest fires have been a major hazard to forest management, needing sophisticated monitoring and management techniques. By creating an embedded intelligent video analysis system, this research proposed a complete strategy for addressing this difficulty. Method:: The system's hardware architecture was explained, and the operating system software was detailed, using a software and hardware design based on the ZynqSoC. At the same time, an emphasis on forest fire prevention applications was maintained. Furthermore, the study investigated a unique technique for forest fire detection using Arduino as a field data collector and a fuzzy logic algorithm to improve accuracy. Results:: The proposed IoT-Fog-Cloud collaboration infrastructure offered a patented contribution to real-time wildfire monitoring, prediction, and forecasting. The framework achieved excellent accuracy in determining wildfire proneness levels and real-time alert production by utilizing fuzzy K-nearest-neighbor classification and Holt-Winter's forecasting model. Conclusion:: The findings demonstrated the integrated system's ability to reduce the impact of wildfires, serving as a significant reference for future forest fire prevention scenarios.

Reliability and Validity of Temporal Parameters during the Side Hop Test in Patients after Anterior Cruciate Ligament Reconstruction.

Background: The side hop test (SHT) measures the number of jumps performed over 30 s. Although this measure has demonstrated its value in clinical practice, the temporal parameters of the SHT allow for a deeper analysis of the execution strategy. The aim of this study is to assess the reliability and construct validity of contact time parameters during the SHT recorded by a video analysis system in anterior cruciate ligament reconstructed (ACLR) patients. Methods: We investigated the reliability (intra-rater, standard error of measurement (SEM), and minimum detectable change (MDC)), discriminant validity (operated (OP) versus non-operated (NOP) side), and convergent validity (relationship with strength and psychological readiness) of SHT contact time parameters, number of valid hops and limb symmetry index (LSI) in 38 ACLR patients. Contact time parameters are presented as mean, standard deviation (SD), and coefficient of variation (CV) of contact time. Results: Intra-tester reliability was good to excellent for all contact time parameters. For discriminant validity, the mean and SD contact times of the OP leg were significantly longer than those of the NOP leg, although the difference was smaller than the SEM and MDC values. The number of valid jumps and CV contact time parameters were not significantly different. Isokinetic quadriceps strength (60°/s) was strongly correlated with mean contact time for both legs. However, psychological readiness was not correlated with any of the contact time parameters. Conclusions: Temporal parameters of the SHT measured on video analysis are valid and reliable parameters to assess the performance strategy of the SHT. The results should be interpreted with caution regarding the SEM and MDC values. Further studies are needed to measure criterion validity, inter-rater reliability, and responsiveness.

Endothelial function in healthy centenarians living in the Madonie’s district (Italy)

BackgroundEndothelial function declines with age and plays a critical role in cardiovascular health. Therefore, investigating endothelial function in successful aging models, such as centenarians, is of interest. Flow-mediated dilation (FMD) of the brachial artery is the gold standard for measuring endothelial function in vivo in humans. Therefore, we investigated, for the first time, the FMD of the brachial artery in a group of healthy centenarians. MethodsSelected as part of the ABCD project (nutrition, cardiovascular wellness, and diabetes) centenarians (aged ≥100 years) living in the municipalities of Madonie (Palermo, Italy) were compared with a younger (aged <65 years) sex-matched control group from the ABCD general cohort. FMD of the brachial artery was measured in all participants using a real-time computed video analysis system for B-mode ultrasound images. Body composition (bioimpedance), carotid intima-media thickness (IMT), and ankle-brachial index (ABI) were also measured. ResultsEleven participants (males 36.4 %; age: 101 ± 1 years) out of 28 healthy centenarians successfully cooperated with the FMD test procedures, which require remaining with the upper limb immobile for approximately 10 min. This subgroup was compared with a control group of 76 healthy and younger individuals (males 36.8 %; aged: 41 ± 14 years; P < 0.001). Centenarians exhibited better endothelial function than the control group (FMD: 12.1 ± 4.3 vs 8.6 ± 5.3 %; P < 0.05). The carotid IMT was higher in the centenarian group than in the control group (0.89 ± 0.09 vs 0.56 ± 0.18 mm; P < 0.001), whereas the ABI was comparable between the two groups. ConclusionsThis small group of centenarians demonstrated an unusually favorable endothelial function, which may contribute to their unique aging profile. Further research is needed to determine whether FMD is a valid prognostic marker for successful aging.

Kinematics of 90° change of direction in young football players: Insights for ACL injury prevention from the CUTtheACLstudy on 6008 trials.

To investigate the 90° change of direction (COD) task in an extensive cohort of competitive healthy football players within the CUTtheACLstudy and to provide normative values and differences between males and females for full-body kinematics based on two-dimensional (2D) video analysis and scoring system. One-thousand-and-two competitive football (soccer) players (age 16.3 ± 2.8 years, 264 females) were prospectively enroled. Each player performed three preplanned 90° COD tasks per limb. The 2D evaluation was performed through objective measures (collected through three high-speed cameras) of frontal and sagittal plane joint kinematics at the cut initial foot contact (IC) and maximum knee flexion angle. A previously published scoring system was adopted to measure the movement quality of the COD task. The scoring system included five criteria (limb stability [LS], pelvis stability [PS], trunk stability [TS], shock absorption [SA], movement strategy [MS]) ranked from 0/2 (nonadequate) to 2/2 (adequate) with a maximum score of 10/10. Normative data were provided for all the variables; statistical differences between male and female players were investigated (p < 0.05). A total of 6008 valid attempts were included. Frontal plane knee projection angle (FPKPA) at initial contact was 24.4 ± 9.8° (95thpercentile: FPKPA > 40°). The total score was ≤4/10 in 71.2% of the trials, the lowest subscores were LS and PS. Female players showed different movement patterns with lower hip and trunk flexion both at IC and maximum knee flexion angle (p < 0.01, ES = 0.41-0.64). Female players also showed worse scores than males in SA, MSand total score (p < 0.01). Female players seem more prone to stiffer lower limb strategy and greater pelvis-trunk frontal plane instability than males. Clinicians could adopt normative data and sex-specific differences in players' movement techniques to improve ACL injury risk mitigation protocols. Level IV.

An Automated Video Analysis System for Retrospective Assessment and Real-Time Monitoring of Endoscopic Procedures (with Video).

Accurate recognition of endoscopic instruments facilitates quantitative evaluation and quality control of endoscopic procedures. However, no relevant research has been reported. In this study, we aimed to develop a computer-assisted system, EndoAdd, for automated endoscopic surgical video analysis based on our dataset of endoscopic instrument images. Large training and validation datasets containing 45,143 images of 10 different endoscopic instruments and a test dataset of 18,375 images collected from several medical centers were used in this research. Annotated image frames were used to train the state-of-the-art object detection model, YOLO-v5, to identify the instruments. Based on the frame-level prediction results, we further developed a hidden Markov model to perform video analysis and generate heatmaps to summarize the videos. EndoAdd achieved high accuracy (>97%) on the test dataset for all 10 endoscopic instrument types. The mean average accuracy, precision, recall, and F1-score were 99.1%, 92.0%, 88.8%, and 89.3%, respectively. The area under the curve values exceeded 0.94 for all instrument types. Heatmaps of endoscopic procedures were generated for both retrospective and real-time analyses. We successfully developed an automated endoscopic video analysis system, EndoAdd, which supports retrospective assessment and real-time monitoring. It can be used for data analysis and quality control of endoscopic procedures in clinical practice.

Wireless Power Transfer System for the Intelligent Video Monitoring and Analysis Based on the Safety of Power Grid Construction Site

This study describes a power grid construction site surveillance system that includes wireless power transmission and an improved Yolo V3 detection model. An artificial intelligence and wireless power transfer methodology is employed used for continuous monitoring. Improved Yolo V3 detection is optimized using Hierarchical Particle Swarm Optimization (HPSO) for enhanced efficacy in detecting power grid development threats. The methodology is compared to existing methods in terms of precision, accuracy, recall, and speed. The findings indicate that the proposed method is achieved a real-time and accurate power grid development site monitoring. Experiments on test sets in various situations improved the detection accuracy, and indicate that the proposed technique is quite robust. Also the detection accuracy and speed of the proposed method have improved over existing approaches, indicate that the strategy provides superior detection efficacy. This method considerably improves risk source detection, that's critical for assuring safety on power grid building sites.

Kinematic analysis of an unrestrained passenger in an autonomous vehicle during emergency braking.

Analyzing human body movement is a critical aspect of biomechanical studies in road safety. While most studies have traditionally focused on assessing the head-neck system due to the restraint provided by seat belts, it is essential to examine the entire pelvis-thorax-head kinematic chain when these body regions are free to move. The absence of restraint systems is prevalent in public transport and is also being considered for future integration into autonomous vehicles operating at low speeds. This article presents an experimental study examining the movement of the pelvis, thorax and head of 18 passengers seated without seat belts during emergency braking in an autonomous bus. The movement was recorded using a video analysis system capturing 100 frames per second. Reflective markers were placed on the knee, pelvis, lumbar region, thorax, neck and head, enabling precise measurement of the movement of each body segment and the joints of the lumbar and cervical spine. Various kinematic variables, including angles, displacements, angular velocities and accelerations, were measured. The results delineate distinct phases of body movement during braking and elucidate the coordination and sequentiality of pelvis, thorax and head rotation. Additionally, the study reveals correlations between pelvic rotation, lumbar flexion, and vertical trunk movement, shedding light on their potential impact on neck compression. Notably, it is observed that the elevation of the C7 vertebra is more closely linked to pelvic tilt than lumbar flexion. Furthermore, the study identifies that the maximum angular acceleration of the head and the maximum tangential force occur during the trunk's rebound against the seatback once the vehicle comes to a complete stop. However, these forces are found to be insufficient to cause neck injury. While this study serves as a preliminary investigation, its findings underscore the need to incorporate complete trunk kinematics, particularly of the pelvis, into braking and impact studies, rather than solely focusing on the head-neck system, as is common in most research endeavors.

Feasibility study of wearable capillary refill time measurement device

AbstractCapillary refill time (CRT) is an internationally accepted indicator of peripheral circulation. The CRT is measured by applying compression to the fingernail for a few seconds, releasing it, and observing the process of refilling of blood at the fingertip. The international guidelines for the management of sepsis and septic shock 2021 contains an additional new recommendation for CRT measurement to determine the peripheral circulation statuses of adult patients with septic shock. However, the current CRT measurement method lacks objectivity. Previous studies have reported the development of measurement devices and video analysis systems, which involve complex measurement environment construction and assume only in-hospital measurements. In addition, since medical workers are limited in the number of carrying medical devices, devices that can be used to obtain multiple biometric indicators in a single measurement are needed. Hence, a prototype wearable CRT measurement device was developed in this study, and its feasibility was evaluated by comparing the agreement, intra-class correlation coefficient, and coefficient of variation with those of a CRT measurement device developed in the past. The results indicated that mean CRTs between measurement methods were agreement, with moderate or better intra-rater reliability and no difference in coefficient of variation. Therefore, the results indicate the feasibility of the proposed wearable CRT device.

РАЗРАБОТКА ПРОГРАММНОГО ОБЕСПЕЧЕНИЯ ДЛЯ АВТОМАТИЗИРОВАННОЙ СИСТЕМЫ ОБРАБОТКИ ВИДЕОМАТЕРИАЛОВ

Trunk pipeline transportation is the most important component of fuel and energy complex of Russia. In accordance with the regulations, control of equipment operation, compliance with labor protection and industrial safety requirements at production facilities as high-risk areas is carried out. Based on the results of monitoring the operation of road construction equipment, it was established that the majority of violations occur in the operation of video recording equipment. To optimize the process of managing video information, the authors have developed a software package (PC) using artificial intelligence methods for digital processing of video content, which allows analyzing data received from video recorders. The software product is designed to automate the process of processing video data: identifying violations in the operation of video recording; operational control and analysis of performance indicators of the DVR; creating a database with identified violations; generating operational, reliable information online from construction equipment. The developed computer program examines video analytics scenarios that make it possible to automate the processing of video materials: defocus or contamination of the video camera lens (cloudy, unclear image, dirty or obscured camera lens); disconnecting the video camera from the DVR (recording is in progress, a black or blue screen is recorded with time stamps on the video); lack of movement (recording of “garbage” video without production of work, no movement, picture does not change); interference when recording video (stripes, strong video jerking, video “scattering” into rectangular segments). The software package consists of 3 main modules, each of which has its own functionality for a comprehensive solution to the task. The use of a software package allows you to identify violations in the operation of video surveillance equipment. Automated analysis of video information is considered as a resource for increasing the efficiency of management and safe operation of special equipment. Using video analysis systems, you can obtain valuable information about the quality of work of company personnel at remote sites, which allows for a more adequate assessment of their work, as well as checking compliance with safety standards at high-risk sites. Implementation of computer vision-based solutions in the business processes of oil companies will allow to achieve higher productivity, profitability and competitiveness of the economic sector.