Activity Detection System Research Articles

Establishment of a new protein C detection system based on chromogenic substrate assay and its clinical diagnostic value for deep vein thrombosis.

BackgroundDeficiency of protein C (PC) affects the balance between blood coagulation and fibrinolysis in the human body. Chromogenic‐based assay is recommended as the preferred screening method for detecting PC deficiency. We established a PC detection system based on the chromogenic substrate assay.MethodsFirst, a kit for the determination of PC activity in plasma was elaborately developed and its reaction parameters on XL‐3200c were explored. Then, we evaluated its performance and collected specimens to compare the test results obtained with those of the Siemens detection system. Finally, the clinical diagnostic efficacy of this detection system for deep vein thrombosis (DVT) was assessed.ResultsOptimum conditions for PC detection were 0.25–0.1 U/ml protein C activator Protac® and 2.5–1 mM Pefachrome®PCa5297. The composition and concentration ranges of buffer substances and stabilizers in the kit were also explored. Satisfactory results were observed in performance evaluation. The test results of the newly built detection system were highly correlated with those of the Siemens detection system (R 2 = 0.9771 in the control group and R 2 = 0.9776 in the DVT group), and Bland‐Altman plots also showed high consistency between the two detection systems. In addition, the area under the curve (AUC) of the newly built PC detection system for DVT was 0.888, indicating this system could effectively improve the diagnostic sensitivity and specificity for DVT.ConclusionIn this study, a sensitive, wide linear range and reliable PC activity detection system were established.

Frog calling activity detection using lightweight CNN with multi-view spectrogram: A case study on Kroombit tinker frog

Frogs play an important role in ecological systems, while frog species across the globe are threatened and declining. Therefore, it is valuable to estimate the frog population based on an intelligent computer system. Due to the success of deep learning (DL) in various pattern recognition tasks, previous studies have used DL-based methods for frog call analysis. However, the performance of DL-based systems is highly affected by their input (feature representation). In this study, we develop a frog calling activity detection system for continuous field recordings using a light convolutional neural network (CNN) with multi-view spectrograms. To be specific, a sliding window is first applied to continuous recordings for obtaining audio segments with a fixed duration. Then, the background noise is filtered out. Next, a multi-view spectrogram is used for characterizing those segments, which has more distinctive information than a single-view spectrogram. Finally, a lightweight CNN model is used for the detection of frog calling activity with a twin loss, where different train and test sets are used to validate the model’s robustness. Our experimental results indicate that the highest macro F1-score was 99.6 ± 0.2 and 96.4 ± 2.0 using 2016 and 2017 as the train data respectively, where CNN-GAP is used as the model with multi-view spectrogram as the input.

Discovery of Camellia sinensis catechins as SARS-CoV-2 3CL protease inhibitors through molecular docking, intra and extra cellular assays

Background and purposePrevious studies suggest that major Camellia sinensis (tea) catechins can inhibit 3-chymotrypsin-like cysteine protease (3CLpro), inspiring us to study 3CLpro inhibition of the recently discovered catechins from tea by our group. MethodsAutodock was used to dock 3CLpro and 16 tea catechins. Further, a 3CLpro activity detection system was used to test their intra and extra cellular 3CLpro inhibitory activity. Surface plasmon resonance (SPR) was used to analyze the dissociation constant (KD) between the catechins and 3CLpro. ResultsDocking data suggested that 3CLpro interacted with the selected 16 catechins with low binding energy through the key amino acid residues Thr24, Thr26, Asn142, Gly143, His163, and Gln189. The selected catechins other than zijuanin D (3) and (-)-8-(5′'R)-N-ethyl-2-pyrrolidinone-3-O-cinnamoylepicatechin (11) can inhibit 3CLpro intracellularly. The extracellular 3CLpro IC50 values of (–)-epicatechin 3-O-caffeoate (EC-C, 1), zijuanin C (2), etc-pyrrolidinone C and D (6), etc-pyrrolidinone A (9), (+)-gallocatechin gallate (GCG), and (-)-epicatechin gallate (ECG) are 1.58 ± 0.21, 41.2 ± 3.56, 0.90 ± 0.03, 46.71 ± 10.50, 3.38 ± 0.48, and 71.78 ± 8.36 µM, respectively. The KD values of 1, 6, and GCG are 4.29, 3.46, and 3.36 µM, respectively. ConclusionTogether, EC-C (1), etc-pyrrolidinone C and D (6), and GCG are strong 3CLpro inhibitors. Our results suggest that structural modification of catechins could be conducted by esterificating the 3-OH as well as changing the configuration of C-3, C-3′'' or C-5′'' to discover strong SARS-CoV-2 inhibitors.

Effect of energy renovation and occupants' activities on airborne particle concentrations in Swedish rental apartments

Exposure to airborne particles causes detrimental health effects, hence their assessment in indoor environments, where people spend most of the time, is important. The influence of energy renovation and occupants' activities on airborne particle concentrations was assessed in seven occupied Swedish residences. Ultrafine particle (UFP, <100 nm) number concentrations, PM2.5 (<2.5 μm) and black carbon (BC, <900 nm) mass concentrations were simultaneously measured inside and outside before, after renovation, and during follow-up.The average indoor UFP number concentrations increased from 6200 (±4070) cm-3 before renovation to 12,700 (±6040) cm-3 during the follow up, as the number of indoor activities doubled. Indoor UFP number concentrations depended mainly on frequency and type of occupants' activities in studied residences (e.g., cooking, candle burning). The average indoor PM2.5 concentration decreased from 8.6 (±5.8) μg m-3 before renovation to 2.5 (±1.3) μg m-3 during follow up, as the activities that generated PM2.5 decreased, and infiltration of outdoor particles could have been decreased due to renovation measures. However, the indication of infiltration decrease during the follow up, assessed on the basis of indoor to outdoor ratios during non-activity times (with no influence of occupants' activities), was not observed after the renovation and should be treated with caution. In this study indoor PM2.5 and BC were influenced by activities and outdoor concentrations. Reduction of exposure to indoor UFP, might be obtained by optimization of kitchen exhaust flows. An improved design of supply air inlets in mechanical exhaust ventilation systems may reduce PM2.5 infiltration.Occupants' logbook records, needed for identification of sources contributing to particle exposure, proved useful but not always accurate compared to technical measurements of activities and UFP concentrations. Development of simple and reliable activity detection systems is needed to complement logbooks and allow accurate assessment of source contribution to particle exposure in homes and associated health effects.

HuRAI: A brain-inspired computational model for human-robot auditory interface

The deep learning era endows immense opportunities for ubiquitous robotic applications by leveraging big data generated from widespread sensors and ever-growing computing capability. While the growing demands for natural human-robot interaction (HRI) as well as concerns for energy efficiency, real-time performance, and data security motive novel solutions. In this paper, we present a brain-inspired spiking neural network (SNN) based Human-Robot Auditory Interface, namely HuRAI. The HuRAI integrates the voice activity detection, speaker localization and voice command recognition systems into a unified framework that can be implemented on the emerging low-power neuromorphic computing (NC) devices. Our experimental results demonstrate superior modeling capabilities of SNNs, achieving accurate and rapid prediction for each task. Moreover, the energy efficiency analysis reveals a compelling prospect, with up to three orders of magnitude energy savings, over the equivalent artificial neural networks that running on the state-of-the-art Nvidia graphics processing unit (GPU). Therefore, integrating the algorithmic power of large-scale SNN models and the energy efficiency of NC devices offers an attractive solution for real-time, low-power robotic applications.

Designing a Computer-Vision Application: A Case Study for Hand-Hygiene Assessment in an Open-Room Environment.

Hand-hygiene is a critical component for safe food handling. In this paper, we apply an iterative engineering process to design a hand-hygiene action detection system to improve food-handling safety. We demonstrate the feasibility of a baseline RGB-only convolutional neural network (CNN) in the restricted case of a single scenario; however, since this baseline system performs poorly across scenarios, we also demonstrate the application of two methods to explore potential reasons for its poor performance. This leads to the development of our hierarchical system that incorporates a variety of modalities (RGB, optical flow, hand masks, and human skeleton joints) for recognizing subsets of hand-hygiene actions. Using hand-washing video recorded from several locations in a commercial kitchen, we demonstrate the effectiveness of our system for detecting hand hygiene actions in untrimmed videos. In addition, we discuss recommendations for designing a computer vision system for a real application.

MC-LSTM: Real-Time 3D Human Action Detection System for Intelligent Healthcare Applications.

Due to the movement expressiveness and privacy assurance of human skeleton data, 3D skeleton-based action inference is becoming popular in healthcare applications. These scenarios call for more advanced performance in application-specific algorithms and efficient hardware support. Warnings on health emergencies sensitive to response speed require low latency output and action early detection capabilities. Medical monitoring that works in an always-on edge platform needs the system processor to have extreme energy efficiency. Therefore, in this paper, we propose the MC-LSTM, a functional and versatile 3D skeleton-based action detection system, for the above demands. Our system achieves state-of-the-art accuracy on trimmed and untrimmed cases of general-purpose and medical-specific datasets with early-detection features. Further, the MC-LSTM accelerator supports parallel inference on up to 64 input channels. The implementation on Xilinx ZCU104 reaches a throughput of 18658 Frames-Per-Second (FPS) and an inference latency of 3.5ms with the batch size of 64. Accordingly, the power consumption is 3.6W for the whole FPGA+ARM system, which is 37.8x and 10.4x more energy-efficient than the high-end Titan X GPU and i7-9700 CPU, respectively. Meanwhile, our accelerator also keeps a 4 ∼5x energy efficiency advantage against the low-power high-performance Firefly-RK3399 board carrying an ARM Cortex-A72+A53 CPU. We further synthesize an 8-bit quantized version on the same hardware, providing a 48.8% increase in energy efficiency under the same throughput.

Deep-violence: individual person violent activity detection in video

The need for automatic activity detection systems has been elevated since the number of surveillance cameras installed in the surroundings is increased. Automatic activity detection systems can be productively used to cooperate with human operators and for offline inspection to generate an on-line alarm in case of abnormal activities. Although the activity detection problem is a trending field belonging to computer vision, automatically characterizing violent scenes has been considerably less studied in the surveillance system which is vindicated by the demand of providing safer surroundings for the public. Thus in the proffered work, a deep neural network model based on an ensemble of the Mask Region-based Convolutional Neural network (Mask RCNN), Key-point detection, and Long Short Term Memory (LSTM) has been put forward to identify single person, violent activities such as Punching, kicking. Former to extract human key-points and mask and later to capture temporal information of the data. The upshot of experiments manifests that the ensemble model can outperform individual models. The proposed approach has managed to accomplish a good accuracy rate of 73.1%, 93.4%, and 86.5% on Weizmann, KTH, and own Dataset respectively. The proposed work is more relevant to the industry, which in turn is helpful in serving society as it deals with security.

Detection of individual activities in video sequences based on fast interference discovery and semi-supervised method

Auto understanding of human activities in video is an increasing necessity in some application realms. The existing methods for human’s activity identification are divided into two methods: activity recognition and activity detection. The most important challenge in activity detection realm is activity boundary false detection which decreases system accuracy. In this research, an activity detection system was suggested denoting rapid interference and sewing it. Although it has improved accuracy it has also accuracy time, activities in suggested system were replayed more usefully and influenced by creating a descriptor denoting movable and apparent form. The suggested system was tested on Weizmann dataset and reached an accuracy of 93.34%. Furthermore, the proposed system in activity recognition was tested on KTH dataset and reached an accuracy of 93.63%. When activity recognition is stated as a learning case, sufficient labeled educational examples must be used. But labeling the video data is expensive, so the useful method uses unlabeled and labeled examples, during the learning process, this idea is the basic foundation of the semi-supervised method. In this research, a semi-supervised method with co-training algorithm appearance and active learning was suggested which improved the efficiency of semi-supervised learning that was tested.

Promoting Data Availability Framework by Using Gamification on Smartphone Fall Detection Based Human Activities

Abstract Researcher usually can get Activities of Daily Living (ADL) data without extra effort because those activities normally performed in daily life. Researcher usually faced a problem if want to develop a fall detection because fall can be categorizing as abnormal activity, it is an incident and occurs rarely. Fall data collected in experimental environment, which is not be able to represent the actual pattern data. Volunteers also can made up a movement, so that the fall movement becomes unnatural. The advantages of gamification are adopted by us to create a solution of data availability in human activities data especially for fall detection. Gamification using badge adopted in the activity detection system to detect ADL and falls can help the researcher to collect the data because user will be motivated to collect the badges. Promoting data availability framework on smartphone fall detection based human activities that consist of three main components: activity detection method (ADL and fall detection method), badges, and data-activity can be a guidance to help researcher to collect actual fall data and ADL for future research (develop a better ADL and fall detection).

The Drosophila Individual Activity Monitoring and Detection System (DIAMonDS).

Here, we have developed DIAMonDS (Drosophila Individual Activity Monitoring and Detection System) comprising time-lapse imaging by a charge-coupled device (CCD) flatbed scanner and Sapphire, a novel algorithm and web application. DIAMonDS automatically and sequentially identified the transition time points of multiple life cycle events such as pupariation, eclosion, and death in individual flies at high temporal resolution and on a large scale. DIAMonDS performed simultaneous multiple scans to measure individual deaths (≤1152 flies per scanner) and pupariation and eclosion timings (≤288 flies per scanner) under various chemical exposures, environmental conditions, and genetic backgrounds. DIAMonDS correctly identified 74-85% of the pupariation and eclosion events and ~ 92% of the death events within ± 10 scanning frames. This system is a powerful tool for studying the influences of genetic and environmental factors on fruit flies and efficient, high-throughput genetic and chemical screening in drug discovery.

Automatically Detected Pecking Activity in Group-Housed Turkeys.

Simple SummaryCannibalism is one of the biggest welfare issues of today’s turkey husbandry. We hypothesized that changes in pecking activity might indicate imminent cannibalism. Therefore, in this pilot study a newly developed automatic pecking activity detection was validated, and continuously applied to gain information about pecking activity of group-housed turkeys during the rearing phase and before a cannibalistic outbreak. The pecking object was used by turkeys the whole recording time. Activity on the object was highest in the morning. No clear trend in pecking activity development before an outbreak has yet been found. Pecking detection has to be further tested under various farm conditions. The system can be used in further research in order to survey changes in pecking activity in turkeys.In search for an early warning system for cannibalism, in this study a newly developed automatic pecking activity detection system was validated and used to investigate how pecking activity changes over the rearing phase and before cannibalistic outbreaks. Data were recorded on two farms, one with female (intact beaks) and the other with male (trimmed beaks) turkeys. A metallic pecking object that was equipped with a microphone was installed in the barn and video monitored. Pecking activity was continuously recorded and fed into a CNN (Convolutional neural network) model that automatically detected pecks. The CNN was validated on both farms, and very satisfactory detection performances were reached (mean sensitivity/recall, specificity, accuracy, precision, and F1-score around 90% or higher). The extent of pecking at the object differed between farms, but the objects were used during the whole recording time, with highest activities in the morning hours. Daily pecking frequencies showed a low downward trend over the rearing period, although on both farms they increased again in week 5 of life. No clear associations between pecking frequencies and in total three cannibalistic outbreaks on farm 1 in one batch could be found. The detection system is usable for further research, but it should be further automated. It should also be further tested under various farm conditions.

Orientation Invariant Skeleton Feature (OISF): a new feature for Human Activity Recognition

Human Activity Recognition is the process of identifying the activity of a person by analyzing continuous frames of a video. In many application areas, human activity identification is either a direct goal or it is a key segment of a bigger objective. Some of the examples are surveillance system, elder healthcare monitoring system, abnormal activity detection systems such as fight detection, theft detection etc. Robust and accurate activity recognition is a challenging task due to diverse reasons, such as changing ambient illumination, noise, background turbulence, camera placements etc. Existing literatures discuss some techniques for identifying human activity but these approaches are restricted to the case of videos recorded from static camera. The aim of the proposed approach is to fill this gap. In this proposed method, a new skeleton based feature for human activity recognition- “Orientation Invariant Skeleton Feature (OISF)”- is introduced and used to train Random Forest (RF) classifier for Human Activity Recognition. Efficiency of newly introduced feature OISF is analyzed for the videos recorded with multiple cameras positioned at two different slant angles. Experimental results reveal that the newly introduced feature OISF has minimal dependency on variations of camera orientation. Accuracy achieved is ≈ 99.30% with ViHASi dataset, ≈ 96.85% with KTH dataset and ≈ 98.34% with in-house dataset which is higher than those achieved by other researches with existing features. The improved result of human activity recognition in terms of accuracy proves the appropriateness of the proposed research in being used commercially.

Design and Analysis for Fall Detection System Simplification.

This paper presents a methodology based on multimodal sensors to configure a simple, comfortable and fast fall detection and human activity recognition system that can be easily implemented and adopted. The methodology is based on the configuration of specific types of sensors, machine-learning methods and procedures. The protocol is divided into four phases: (1) database creation (2) data analysis (3) system simplification and (4) evaluation. Using this methodology, we created a multimodal database for fall detection and human activity recognition, namely UP-Fall Detection. It comprises data samples from 17 subjects that perform 5 types of falls and 6 different simple activities, during 3 trials. All information was gathered using 5 wearable sensors (tri-axis accelerometer, gyroscope and light intensity), 1 electroencephalograph helmet, 6 infrared sensors as ambient sensors, and 2 cameras in lateral and front viewpoints. The proposed novel methodology adds some important stages to perform a deep analysis of the following design issues in order to simplify a fall detection system: a) select which sensors or combination of sensors are to be used in a simple fall detection system, b) determine the best placement of the sources of information, and c) select the most suitable machine learning classification method for fall and human activity detection and recognition. Even though some multimodal approaches reported in literature only focus on one or two of the above-mentioned issues, our methodology allows simultaneously solving these three design problems related to a human fall and activity detection and recognition system.

Low-Energy Voice Activity Detection via Energy-Quality Scaling From Data Conversion to Machine Learning

In this work, voice activity detection (VAD) systems with system-level energy-quality (EQ) scaling are investigated. Compared to prior single-knob EQ scaling, multiple EQ knobs are selectively inserted into the entire signal chain from end to end. EQ knobs are dynamically co-optimized to minimize energy for a given quality target. The analysis shows that system-level EQ optimization provides several benefits and has interesting implications on the performance of machine learning-based classification, as exemplified by decision trees in this work. First, it can make quality degradation more graceful than single-knob, allowing for more aggressive energy reduction under a given quality target, while retaining the ability to operate at full quality. Also, proper system-level EQ optimization enhances fitting in machine learning-based systems (e.g., decision tree-based), suppressing both underfitting and overfitting. The analysis also shows that context-specific retraining significantly improves quality and resolves fitting issues, especially at low input SNR. Measurements on a 28nm testchip show that system-level EQ scaling can reduce energy by up to 3.5X at 2% accuracy degradation in 10-dB noise, compared to full quality. Iso-technology comparison shows that the minimum energy of 51.9 nJ/frame is lower than prior art by 1.9-74.4X at comparable speech/non-speech hit rates.

Abnormal Human Activity Detection using Unsupervised Machine Learning Techniques

Nowadays there is a significant study effort due to the popularity of CCTV to enhance analysis methods for surveillance videos and video-based images in conjunction with machine learning techniques for the purpose of independent assessment of such information sources. Although recognition of human intervention in computer vision is extremely attained subject, abnormal behavior detection is lately attracting more research attention. In this paper, we are interested in the studying the two main steps that compose abnormal human activity detection system which are the behavior representation and modelling. And we use different techniques, related to feature extraction and description for behavior representation as well as unsupervised classification methods for behavior modelling. In addition, available datasets and metrics for performance evaluation will be presented. Finally, this paper will be aimed to detect abnormal behaved object in crowd, such as fast motion in a crowd of walking people

A Two-Level Flow-Based Anomalous Activity Detection System for IoT Networks

The significant increase of the Internet of Things (IoT) devices in smart homes and other smart infrastructure, and the recent attacks on these IoT devices, are motivating factors to secure and protect IoT networks. The primary security challenge to develop a methodology to identify a malicious activity correctly and mitigate the impact of such activity promptly. In this paper, we propose a two-level anomalous activity detection model for intrusion detection system in IoT networks. The level-1 model categorizes the network flow as normal flow or abnormal flow, while the level-2 model classifies the category or subcategory of detected malicious activity. When the network flow classified as an anomaly by the level-1 model, then the level-1 model forwards the stream to the level-2 model for further investigation to find the category or subcategory of the detected anomaly. Our proposed model constructed on flow-based features of the IoT network. Flow-based detection methodologies only inspect packet headers to classify the network traffic. Flow-based features extracted from the IoT Botnet dataset and various machine learning algorithms were investigated and tested via different cross-fold validation tests to select the best algorithm. The decision tree classifier yielded the highest predictive results for level-1, and the random forest classifier produced the highest predictive results for level-2. Our proposed model Accuracy, Precision, Recall, and F score for level-1 were measured as 99.99% and 99.90% for level-2. A two-level anomalous activity detection system for IoT networks we proposed will provide a robust framework for the development of malicious activity detection system for IoT networks. It would be of interest to researchers in academia and industry.

Fleet Monitoring System using GSM, GPS and Web Server with Drunken Driving Alerts

This paper explains the design and implementation of an on-demand vehicle tracking and alcohol activity detection system using microcontroller and suitable peripherals. Correct information about a vehicle’s location is extremely important in managing fleet, logistics and cab services. Unfortunately, operators or owners of such services find a tough time to perfectly spot their trucks or vehicles owing to irresponsible, reluctant, untruthful and unlawful drivers. We propose a system based on Global System for Mobile Communication (GSM), Global Positioning System (GPS), PIC microcontroller, alcohol sensor etc., which can provide the whereabouts of the vehicle in the form of an SMS to the owner’s mobile phone. A working prototype of the proposed system was created and tested successfully. Such systems will be of great utility to cab and fleet operators in developing countries such as India etc.

HCF: A Hybrid CNN Framework for Behavior Detection of Distracted Drivers

Distracted driving causes a large number of traffic accident fatalities and is becoming an increasingly important issue in recent research on traffic safety. Gesture patterns are less distinguishable in vehicles due to in-vehicle physical constraints and body occlusions from the drivers. However, by capitalizing on modern camera technology, convolutional neural network (CNN) can be used for visual analysis. In this paper, we present a hybrid CNN framework (HCF) to detect the behaviors of distracted drivers by using deep learning to process image features. To improve the accuracy of the driving activity detection system, we first apply a cooperative pretrained model that combines ResNet50, Inception V3 and Xception to extract driver behavior features based on transfer learning. Second, because the features extracted by pretrained models are independent, we concatenate the extracted features to obtain comprehensive information. Finally, we train the fully connected layers of the HCF to filter out anomalies and hand movements associated with non-distracted driving. We apply an improved dropout algorithm to prevent the proposed HCF from overfitting to the training data. During the evaluation, we apply the class activation mapping (CAM) technique to highlight the feature area involving ten tested classes of typical distracted driving behaviors. The experimental results show that the proposed HCF achieves the classification accuracy of 96.74% when detecting distracted driving behaviors, demonstrating that it can potentially help drivers maintain safe driving habits.

Spectro-Temporal Attention-Based Voice Activity Detection

Voice Activity Detection (VAD) systems suffer from unexpected and non-stationary background noises at magnitudes sufficiently high to mask the speech signal.Although several methods of increasing the performance of VAD have been proposed, their approaches have yet to mitigate the influence of the background noise itself. This letter proposes an effective noise-robust VAD system approach. The proposed method uses spectral attention and temporal attention through applying a deep learning-based attention mechanism. The proposed method is demonstrated and compared with several other deep learning-based methods in terms of the area under the curve in experiments with either known or unknown noise-added, and real-world noisy data. The results show that the proposed method outperforms the other methods in all the scenarios considered, but moreover generalizes well in environments of unknown or unexpected noise.