High Speed and Accuracy of Animation 3D Pose Recognition Based on an Improved Deep Convolution Neural Network

Applied Sciences, Journal Year: 2023, Volume and Issue: 13(13), P. 7566 - 7566

Published: June 27, 2023

Pose recognition in character animations is an important avenue of research computer graphics. However, the current use traditional artificial intelligence algorithms to recognize animation gestures faces hurdles such as low accuracy and speed. Therefore, overcome above problems, this paper proposes a real-time 3D pose system, which includes both facial body poses, based on deep convolutional neural networks further designs single-purpose estimation system. First, we transformed human extracted from input image abstract data structure. Subsequently, generated required at runtime dataset. This challenges conventional concept monocular estimation, extremely difficult achieve. It can also achieve running speed resolution 384 fps. The proposed method was used identify multiple-character using multiple datasets (Microsoft COCO 2014, CMU Panoptic, Human3.6M, JTA). results indicated that improved algorithm performance by approximately 3.5% 8–10 times, respectively, significantly superior other classic algorithms. Furthermore, tested system pose-recognition datasets. attitude reach 24 fps with error 100 mm, considerably less than 2D 60 learning study yielded surprisingly performance, proving deep-learning technology for has great potential.

Language: Английский

---

Enhancing Elderly Fall Detection through IoT-Enabled Smart Flooring and AI for Independent Living Sustainability

Sustainability, Journal Year: 2023, Volume and Issue: 15(22), P. 15695 - 15695

Published: Nov. 7, 2023

In the realm of sustainable IoT and AI applications for well-being elderly individuals living alone in their homes, falls can have severe consequences. These consequences include post-fall complications extended periods immobility on floor. Researchers been exploring various techniques fall detection over past decade, this study introduces an innovative Elder Fall Detection system that harnesses technologies. our configuration, we integrate RFID tags into smart carpets along with readers to identify among population. To simulate events, conducted experiments 13 participants. these experiments, embedded transmit signals readers, effectively distinguishing from events regular movements. When a is detected, activates green signal, triggers alarm, sends notifications alert caregivers or family members. enhance precision detection, employed machine deep learning classifiers, including Random Forest (RF), XGBoost, Gated Recurrent Units (GRUs), Logistic Regression (LGR), K-Nearest Neighbors (KNN), analyze collected dataset. Results show algorithm achieves 43% accuracy rate, GRUs exhibit 44% XGBoost 33% rate. Remarkably, KNN outperforms others exceptional rate 99%. This research aims propose efficient framework significantly contributes enhancing safety overall independently individuals. It aligns principles sustainability applications.

Language: Английский

---

Multi-Sensor Platform for Predictive Air Quality Monitoring

Sensors, Journal Year: 2023, Volume and Issue: 23(11), P. 5139 - 5139

Published: May 28, 2023

Air quality monitoring is a very important aspect of providing safe indoor conditions, and carbon dioxide (CO2) one the pollutants that most affects people’s health. An automatic system able to accurately forecast CO2 concentration can prevent sudden rise in levels through appropriate control heating, ventilation air-conditioning (HVAC) systems, avoiding energy waste ensuring comfort. There are several works literature dedicated air assessment HVAC systems; performance maximisation such systems typically achieved using significant amount data collected over long period time (even months) train algorithm. This be costly may not respond real scenario where habits house occupants or environment conditions change time. To address this problem, an adaptive hardware–software platform was developed, following IoT paradigm, with high level accuracy forecasting trends by analysing only limited window recent data. The tested considering case study residential room used for smart working physical exercise; parameters analysed were occupants’ activity, temperature, humidity room. Three deep-learning algorithms evaluated, best result obtained Long Short-Term Memory network, which features Root Mean Square Error about 10 ppm training days.

Language: Английский

---

Comparative Analysis of the Clustering Quality in Self-Organizing Maps for Human Posture Classification

Sensors, Journal Year: 2023, Volume and Issue: 23(18), P. 7925 - 7925

Published: Sept. 15, 2023

The objective of this article is to develop a methodology for selecting the appropriate number clusters group and identify human postures using neural networks with unsupervised self-organizing maps. Although clustering algorithms have proven effective in recognizing postures, many works are limited testing which data correctly or incorrectly recognized. They often neglect task groups (where corresponds output neurons, i.e., postures) quality assessments. use scores determine frees expert make subjective decisions about enabling learning. Due high dimensionality variability, (referred as labeling) can be difficult time-consuming. In our case, there no manual labeling step. We introduce new score: discriminant score (DS). describe process most suitable activity records captured by RGB-D cameras. Comparative studies on usefulness popular scores—such silhouette coefficient, Dunn index, Calinski–Harabasz Davies–Bouldin DS—for posture classification tasks presented, along graphical illustrations results produced DS. findings show that DS offers good recognition, effectively following postural transitions similarities.

Language: Английский

---

Sitting Posture Recognition Systems: Comprehensive Literature Review and Analysis

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(18), P. 8557 - 8557

Published: Sept. 23, 2024

Sitting posture recognition systems have gained significant attention due to their potential applications in various domains, including healthcare, ergonomics, and human-computer interaction. This paper presents a comprehensive literature review analysis of existing sitting systems. Through an extensive examination relevant research articles conference papers, we identify analyze the underlying technologies, methodologies, datasets, performance metrics, associated with these The encompasses both traditional methods, such as vision-based approaches sensor-based techniques, well emerging technologies machine learning deep algorithms. Additionally, examine challenges, constraints, future trends field Researchers, practitioners, policymakers who want comprehend most recent developments latest technology will find great value this study.

Language: Английский

---

Sleep posture analysis: state-of-the-art and opportunities of wearable technologies from clinical, sensing and intelligent perception perspectives

Elsevier eBooks, Journal Year: 2025, Volume and Issue: unknown, P. 109 - 134

Published: Jan. 1, 2025

Language: Английский

---

IoT System for Real-Time Posture Asymmetry Detection

Sensors, Journal Year: 2023, Volume and Issue: 23(10), P. 4830 - 4830

Published: May 17, 2023

The rise of the Internet Things (IoT) has enabled development measurement systems dedicated to preventing health issues and monitoring conditions in smart homes workplaces. IoT can support people doing computer-based work avoid insurgence common musculoskeletal disorders related persistence incorrect sitting postures during hours. This proposes a low-cost system for posture symmetry generating visual alert warn worker when an asymmetric position is detected. employs four force sensing resistors (FSR) embedded cushion microcontroller-based read-out circuit pressure exerted on chair seat. Java-based software performs real-time sensors’ measurements implements uncertainty-driven asymmetry detection algorithm. shifts from symmetric vice versa generate close pop-up warning message, respectively. In this way, user promptly notified detected invited adjust position. Every shift recorded web database further analysis behavior.

Language: Английский

---

A Cost-Effective Fall-Detection Framework for the Elderly Using Sensor-Based Technologies

Sustainability, Journal Year: 2023, Volume and Issue: 15(5), P. 3982 - 3982

Published: Feb. 22, 2023

Falls are critical events among the elderly living alone in their rooms and can have intense consequences, such as person being left to lie for a long time after fall. Elderly falling is one of serious healthcare issues that been investigated by researchers over decade, several techniques methods proposed detect fall events. To overcome mitigate issues, fall, this project presents low-cost, motion-based technique detecting all In study, we used IRA-E700ST0 pyroelectric infrared sensors (PIR) mounted on walls around or near patient bed horizontal field view regular motions events; PIR along with Arduino Uno falls save collected data SD classification. For collection, 20 persons contributed patients performing When falls, signal different intensity (high) produced, which certainly differs from signals generated due normal motion. A set parameters was extracted during build dataset. system detects event turns green signal, an alarm generated, message sent inform family members caregivers individual. Furthermore, classified dataset using five machine learning (ML) classifiers, namely: random forest (RF), decision tree (DT), support vector (SVM), naïve Bayes (NB), AdaBoost (AB). Our result reveals RF AB algorithms achieved almost 99% accuracy fall-d\detection.

Language: Английский

---

A Comparative Study on the Performance of 64-bit ARM Processors

International Journal of Interactive Mobile Technologies (iJIM), Journal Year: 2023, Volume and Issue: 17(13), P. 94 - 113

Published: July 4, 2023

Mobile devices are playing an important role in our daily lives. Nowadays, mobile not only phones to call and text, but they also smart that enable users do almost any task could be done on a regular PC. At the heart of design smartphones, there lies processor which all development smartphone arena is attributed. Recently, ARM processors among most prominent used devices, embedded systems. This paper conducts experimental comparative study 64-bit terms performance their effect power consumption, CPU temperature, battery temperature. We use number well-known benchmarks evaluate those characteristics three namely, Snapdragon 778G+, Exynos 1280 HiSilicon Kirin 980. Those smartphones equipped with processors. Our results reveal none three-selected was best characteristics; each has superiority amongst others certain dominated by other characteristics.

Language: Английский

---

Deciphering Optimal Radar Ensemble for Advancing Sleep Posture Prediction through Multiview Convolutional Neural Network (MVCNN) Approach Using Spatial Radio Echo Map (SREM)

Sensors, Journal Year: 2024, Volume and Issue: 24(15), P. 5016 - 5016

Published: Aug. 2, 2024

Assessing sleep posture, a critical component in tests, is crucial for understanding an individual's quality and identifying potential disorders. However, monitoring posture has traditionally posed significant challenges due to factors such as low light conditions obstructions like blankets. The use of radar technolsogy could be solution. objective this study identify the optimal quantity placement sensors achieve accurate estimation. We invited 70 participants assume nine different postures under blankets varying thicknesses. This was conducted setting equipped with baseline eight radars-three positioned at headboard five along side. proposed novel technique generating maps, Spatial Radio Echo Map (SREM), designed specifically data fusion across multiple radars. Sleep estimation using Multiview Convolutional Neural Network (MVCNN), which serves overarching framework comparative evaluation various deep feature extractors, including ResNet-50, EfficientNet-50, DenseNet-121, PHResNet-50, Attention-50, Swin Transformer. Among these, DenseNet-121 achieved highest accuracy, scoring 0.534 0.804 nine-class coarse- four-class fine-grained classification, respectively. led further analysis on ensemble For radars head, single left-located proved both essential sufficient, achieving accuracy 0.809. When only one central head used, omitting side retaining three upper-body resulted accuracies 0.779 0.753, established foundation determining sensor configuration application, while also exploring trade-offs between fewer sensors.

Language: Английский

---