1,720,996 research outputs found
A Smart System for Personal Protective Equipment Detection in Industrial Environments Based on Deep Learning at the Edge
No full textReal-time object detection is currently used to automate various tasks in industrial environments. One of the most important tasks is to improve the safety of workers by monitoring the correct use of Personal Protective Equipment (PPE) in dangerous areas. In this context, usually, a monitoring system analyzes the stream of videos from surveillance cameras to assess PPE usage in real time. When a worker not wearing the appropriate PPE is detected, an acoustic or visual alarm is triggered automatically to raise attention and awareness. The solutions proposed so far are mostly cloud-based systems: images from the site are continuously offloaded to the cloud for analysis. This centralized architecture requires significant network bandwidth to transmit the video feeds through an internet connection that must be reliable, as a network outage would disrupt the service. In this work, we propose a system for real-time PPE detection based on video streaming analysis and Deep Neural Network (DNN). We adopt the edge computing model in which the application for image analysis and classification is deployed on an embedded system installed in proximity of the camera and directly connected to it. The system does not require continuous image transmission towards a cloud system, thus ensuring bandwidth efficiency, reliability, and workers' privacy. A prototype of the proposed system is developed exploiting a low-cost commercial embedded system, i.e. a Raspberry PI, equipped with an Intel Neural Compute Stick 2. We tested the system with five different pre-trained convolutional neural networks (CNNs), fine-tuned to detect different PPEs, namely helmets, vests, and gloves. In our experimental evaluation, we first compared the five CNNs in terms of classification performance and inference latency. Then, we deployed each CNN on the real system and evaluated the system's throughput regarding the number of video frames analyzed each second
Evaluation of NFC-Enabled devices for heterogeneous Wearable Biomedical Application
Full text linkBiomedical systems that aim at monitoring parameters over a long period of time will require non-invasive wearable devices. A wide range of applications, including continuous workers monitoring or daily monitoring of patients, is expected to exploit wearable devices, each one characterized by different requirements. Such devices will adopt a design that minimizes discomfort and does not limit users' mobility to allow round-the-clock wearability. In order to ensure a significant reduction in the size of wearable devices, the next generation of this technology will need no batteries or at least very small ones. In this paper we analyze the use of Near Field Communication (NFC) for ultra-low-power communication in wearable devices for biomedical applications. Our goal is to verify whether NFC technology can support the heterogeneous requirements of different biomedical use-cases. For this purpose, we will first provide a review of the current state of the art in NFC-Enabled solutions. Then, we will evaluate NFC capabilities through a set of experiments, using off-the-shelf devices. We will pose particular focus on the energy harvesting capabilities to evaluate the feasibility of designing battery-less devices. Our results show that NFC adoption in different biomedical applications is possible, as they can ensure proper reading frequency and distance. Finally, we further demonstrate this through a proof-of-concept implementation: an NFC-based sensorized glove for work safety that is able to monitor the external temperature in a continuous manner
Performance Evaluation of YOLOv5 on Edge Devices for Personal Protective Equipment Detection
No full textThe use of personal protective equipment (PPE) is essential to strengthen the safety of workers in the workplace. Although there exist specific regulations requiring the use of PPEs, due to carelessness, haste or comfort, workers sometimes neglect to wear them. Thus, it is crucial to monitor the appropriate use of PPE, especially in dangerous processes. Computer vision technology can help perform this task automatically, exploiting appropriate deep neural models for recognizing PPE. YOLO (You Only Look Once) deep neural models ensure good accuracy against a limited complexity, which allows running them on devices with limited computational capacity. In this paper, we evaluate the performances, in terms of accuracy and processing speed, of the most popular models implemented in the version 5 of YOLO (YOLOv5) in the task of recognizing whether workers correctly wear PPE. To this aim, all models are trained on a PPE dataset on a dedicated server. Then, each model is deployed on low-cost hardware, which includes a Raspberry Pi4 Model B equipped with an Intel Neural Compute Stick 2, used as a processing unit. The outcomes show that YOLOv5n, with only 1.9 million parameters, is the fastest model which allows processing 7.9 frames per second, while YOLOv5l and YOLOv5x, respectively with 46.5 and 86.7 million parameters, are the most accurate but slowest models, processing 1.3 and 0.7 frames per second. We also compare the performance of the YOLOv5 models with the ones in version 4 of YOLO, showing how models in version 5 in general outperform the previous version with higher accuracy, especially in the detection of small objects
Botulinum toxin type A in the healing of ulcer following oro-mandibular dyskinesia in a patient in a vegetative state
No full textOBJECTIVE: Use of botulinum toxin is expanding as clinical studies demonstrate new potential therapeutic applications. In rehabilitation, botulinum toxin is predominantly used as adjunct therapy for the treatment of spasticity, but it may prove useful for other atypical clinical situations.
CASE HISTORY: A 73-year-old man had a severe sub-arachnoid haemorrhage following the rupture of a giant aneurism of the middle left cerebral artery. Clinically, the patient presented a vegetative state and an oro-mandibular dyskinesia that produced a chronic ulcer on the lower lip. As treatment for this dyskinesia, a total of 320 U botulinum toxin type A were injected into the upper and lower orbicularis oris and masseter muscles.
RESULTS AND DISCUSSION: This treatment allowed for application of topical medication and subsequently, ulcer healing. Botulinum toxin type A may be an important therapeutic aid for clinicians faced with treating persistent pathological conditions caused by dyskinesi
Does Spasticity Reduction by Botulinum Toxin Type A Improve Upper Limb Functionality in Adult Post-Stroke Patients? A Systematic Review of Relevant Studies.
Full text linkShort-Range Localization via Bluetooth Using Machine Learning Techniques for Industrial Production Monitoring
Full text linkIndoor short-range localization is crucial in many Industry 4.0 applications. Production monitoring for assembly lines, for instance, requires fine-grained positioning for parts or goods in order to keep track of the production process and the stations traversed by each product. Due to the unavailability of the Global Positioning System (GPS) for indoor positioning, a different approach is required. In this paper, we propose a specific design for short-range indoor positioning based on the analysis of the Received Signal Strength Indicator (RSSI) of Bluetooth beacons. To this aim, different machine learning techniques are considered and assessed: regressors, Convolution Neural Network (CNN) and Recurrent Neural Network (RNN). A realistic testbed is created to collect data for the training of the models and to assess the performance of each technique. Our analysis highlights the best models and the most convenient and suitable configuration for indoor localization. Finally, the localization accuracy is calculated in the considered use case, i.e., production monitoring. Our results show that the best performance is obtained using the K-Nearest Neighbors technique, which results in a good performance for general localization and in a high level of accuracy, 99%, for industrial production monitoring
A Smart System for Personal Protective Equipment Detection in Industrial Environments Based on Deep Learning
No full textThe adoption of real-time object detection systems via video streaming analysis is currently exploited in several contexts, from security monitoring to safety prevention. In industrial environments, proper usage of Personal Protective Equipment (PPE) is paramount to ensure workers' safety. However, the use of some types of PPE, such as helmets, is often neglected by workers, especially in indoor areas. Thus, in order to reduce the risks of accidents, real-time video streaming-based monitoring systems may be used to monitor areas in which workers operate and alert them not to wear PPEs via acoustic alarms or visual signals. In case of a remote analysis, there are potential issues related to the high rate of data streams to be transported and analyzed and workers' privacy. In this work, we propose an embedded smart system for real-time PPE detection based on video streaming analysis and deep learning models. We discuss the deployment of different versions of the YOLOv4 network fine-tuned using a public PPE dataset. In the end, we assess the performance of the proposed system in terms of accuracy and latency and of the overall PPE detection procedure
A sensorized glove for industrial safety based on Near-Field Communication
No full textRecent Information Technologies are progressively dramatically changing how industrial processes are carried out. This is leading to a new industrial revolution, Industry 4.0, in which productivity, efficiency and safety are improved significantly. Safety of workers and operators, in particular, is expected to improve thanks to pervasive technologies like wearable devices and sensors. In this paper we present a demo of a sensorized glove for industrial safety based on Near-Field Communication (NFC). The demonstration is a proof-of-concept implementation that shows how an NFC-enabled sensor installed in the glove of a worker can be exploited to monitor worker's action and report dangerous situations in advance in order to prevent accidents and injuries
High Dosage of Botulinum Toxin Type A in Adult Subjects with Spasticity Following Acquired Central Nervous System Damage: Where Are We at?
Full text linkSpasticity is a common disabling disorder in adult subjects suffering from stroke, brain injury, multiple sclerosis (MS) and spinal cord injury (SCI). Spasticity may be a disabling symptom in people during rehabilitation and botulinum toxin type A (BTX-A) has become the first-line therapy for the local form. High BTX-A doses are often used in clinical practice. Advantages and limitations are debated and the evidence is unclear. Therefore, we analysed the efficacy, safety and evidence for BTX-A high doses. Studies published from January 1989 to February 2020 were retrieved from MEDLINE/PubMed, Embase, Cochrane Central Register. Only obabotulinumtoxinA (obaBTX-A), onabotulinumtoxinA (onaBTX-A), and incobotulinumtoxinA (incoBTX-A) were considered. The term "high dosage" indicated ≥ 600 U. Thirteen studies met the inclusion criteria. Studies had variable method designs, sample sizes and aims, with only two randomised controlled trials. IncoBTX-A and onaBTX-A were injected in three and eight studies, respectively. BTX-A high doses were used predominantly in treating post-stroke spasticity. No studies were retrieved regarding treating spasticity in MS and SCI. Dosage of BTX-A up to 840 U resulted efficacious and safety without no serious adverse events (AEs). Evidence is insufficient to recommend high BTX-A use in clinical practice, but in selected patients, the benefits of high dose BTX-A may be clinically acceptable
- …
