• International Journal of Computer Science & Network Security
• /
• 제21권8호
• /
• pp.127-136
• /
• 2021

Real-time human recognition is a challenging task, as the images are captured in an unconstrained environment with different poses, makeups, and styles. This limitation is addressed by generating several facial images with poses, makeup, and styles with a single reference image of a person using Generative Adversarial Networks (GAN). In this paper, we propose deep learning-based human recognition using integration of GAN and Spatial Domain Techniques. A novel concept of human recognition based on face depiction approach by generating several dissimilar face images from single reference face image using Domain Transfer Generative Adversarial Networks (DT-GAN) combined with feature extraction techniques such as Local Binary Pattern (LBP) and Histogram is deliberated. The Euclidean Distance (ED) is used in the matching section for comparison of features to test the performance of the method. A database of millions of people with a single reference face image per person, instead of multiple reference face images, is created and saved on the centralized server, which helps to reduce memory load on the centralized server. It is noticed that the recognition accuracy is 100% for smaller size datasets and a little less accuracy for larger size datasets and also, results are compared with present methods to show the superiority of proposed method.

• Journal of Information Processing Systems
• /
• 제15권4호
• /
• pp.833-852
• /
• 2019

In real time applications, due to their effective cost and small size, wireless networks play an important role in receiving particular data and transmitting it to a base station for analysis, a process that can be easily deployed. Due to various internal and external factors, networks can change dynamically, which impacts the localisation of nodes, delays, routing mechanisms, geographical coverage, cross-layer design, the quality of links, fault detection, and quality of service, among others. Conventional methods were programmed, for static networks which made it difficult for networks to respond dynamically. Here, machine learning strategies can be applied for dynamic networks effecting self-learning and developing tools to react quickly and efficiently, with less human intervention and reprogramming. In this paper, we present a wireless networks survey based on different machine learning algorithms and network lifetime parameters, and include the advantages and drawbacks of such a system. Furthermore, we present learning algorithms and techniques for congestion, synchronisation, energy harvesting, and for scheduling mobile sinks. Finally, we present a statistical evaluation of the survey, the motive for choosing specific techniques to deal with wireless network problems, and a brief discussion on the challenges inherent in this area of research.

• 산업진흥연구
• /
• 제8권4호
• /
• pp.123-129
• /
• 2023

본 연구는 초고령사회에서 겪을 수 있는 노인에 대한 경제적 배제를 논의하고 이를 극복하기 위해 사회관계망 확충을 통해 노인의 인간관계만족도를 향상시키는 방안을 제안하고자 하였다. 본 연구 결과, 첫째, 경제적 배제를 극복하고 인간만족도를 향상시키는 방법은 그 대상에 있어서 여성에 대한 관심이 높아야 된다는 것을 암시하고 있으며, 고연령층, 저학력층, 저소득층에 대한 경제적 배제의 극복이 시급한 것으로 나타났다. 둘째, 사회관계망이 인간관계만족도에 미치는 영향을 조사한 결과, 여성일수록, 연령이 높을수록, 주소비처가 쇼핑일수록, 자녀와의 소통이 원활할수록 인간관계 만족도가 높아지는 것으로 나타났다. 따라서 사회관계에 대한 개선책이 요구된다고 볼 수 있다. 셋째, 경제적 배제가 인간관계 만족도에 미치는 영향을 조사한 결과, 여가활동이 친지 및 친척 방문일수록, 사회관계망 이용처가 유료시설일수록 인간관계 만족도가 낮아지는 것으로 나타났다. 따라서 여가활동과 사회관계망에 대한 개선책이 요구된다고 볼 수 있다. 넷째, 사회관계망의 매개효과를 조사한 결과, 소득영역에서의 배제, 노동시장의 배제, 주거복지의 배제를 항목으로 하는 독립변수인 경제적 배제가 종속변수인 인간관계만족도에 영향을 미치는 인과관계에서 사회관계망은 완전 매개효과가 있는 것으로 나타났다. 결론적으로 경제적 배제와 사회관계망은 인간관계만족에 영향을 끼치며, 경제적 배제가 극복되어 사회관계망이 개선될 때 비로소 인간관계만족도는 향상되는 것으로 나타났다.

• 한국지능시스템학회논문지
• /
• 제15권5호
• /
• pp.632-638
• /
• 2005

This paper describes a new data-handing, based on a Spatial Human Interface as human indicator, to the Spatial-Knowledge-Tags (SKT) in the spatial memory the Spatial Human Interface (SHI) is a new system that enables us to facilitate human activity in a working environment. The SHI stores human activity data as knowledge and activity history of human into the Spatial Memory in a working environment as three-dimensional space where one acts, and loads them with the Spatial-Knowledge-Tags(SKT) by supporting the enhancement of human activity. To realize this, the purpose of SHI is to construct new relationship among human and distributed networks computers and sensors that is based on intuitive and simultaneous interactions. In this paper, the specified functions of SKT and the realization method of SKT are explained. The utility of SKT is demonstrated in designing a robot motion control.

• 한국인터넷방송통신학회논문지
• /
• 제22권5호
• /
• pp.177-183
• /
• 2022

최근 COVID-19로 인해 홈 트레이닝의 관심도가 증가하고 있다. 이에 따라 HAR(human activity recognition) 기술을 홈 트레이닝에 적용한 연구가 진행되고 있다. 기존 HAR 분야의 논문에서는 동적인 자세보다는 앉기, 일어서기와 같은 정적인 자세들을 분석한다. 본 논문은 동적인 운동 자세를 분석하여 사용자의 운동 자세 정확도를 보여주는 딥러닝 모델을 제안한다. AI hub의 피트니스 이미지를 blaze pose를 사용하여 사람의 자세 데이터를 분석한다. 3개의 딥러닝 모델: RNN(recurrnet neural networks), LSTM(long short-term memory networks), CNN(convolution neural networks)에 대하여 실험을 진행한다. RNN, LSTM, CNN 모델의 f1-score는 각각 0.49, 0.87, 0.98로 CNN 모델이 가장 적합하다는 것을 확인하였다. 이후 연구로는, 다양한 학습 데이터를 사용하여 더 많은 운동 자세를 분석할 예정이다.

• Journal of Electrical Engineering and Technology
• /
• 제10권4호
• /
• pp.1910-1914
• /
• 2015

Human posture recognition is an attractive and challenging topic in computer vision due to its promising applications in the areas of personal health care, environmental awareness, human-computer-interaction and surveillance systems. Human posture recognition in video sequences consists of two stages: the first stage is training and evaluation and the second is deployment. In the first stage, the system is trained and evaluated using datasets of human postures to ‘teach’ the system to classify human postures for any future inputs. When the training and evaluation process is deemed satisfactory as measured by recognition rates, the trained system is then deployed to recognize human postures in any input video sequence. Different classifiers were used in the training such as Multilayer Perceptron Feedforward Neural networks, Self-Organizing Maps, Fuzzy C Means and K Means. Results show that supervised learning classifiers tend to perform better than unsupervised classifiers for the case of human posture recognition.

• Bioinformatics and Biosystems
• /
• 제2권1호
• /
• pp.24-27
• /
• 2007

Human cell regulatory mechanism is one of suspicious problems among biologists. Here we tried to uncover the human breast cancer cell regulatory mechanism from gene expression data (Marc J. Van de vijver, et. al., 2002) using a module network algorithm which is suggested by Segal, et. al.(2003) Finally, we derived a module network which consists of 50 modules and 10 tree depths. Moreover, to validate this candidate network, we applied a GO enrichment test and known transcription factor-target relationships from Transfac(R) (V. Matys, et. al, 2006) and HPRD database (Peri, S. et al., 2003).

• IEIE Transactions on Smart Processing and Computing
• /
• 제5권3호
• /
• pp.185-192
• /
• 2016

Today, the Internet has penetrated almost all the ins and outs of social life, has changed work, communications, social influence and the lifestyle of humankind. However, it is still short of flexibility, transparency etc., due to network address translator overuse, masschanges, uncomfortable protocols, and so on. Hence, more research is necessary into future telecommunications networks based on contemporary networks accompanied by new requisitions and new designs that are compatible with today's and tomorrow's demands. This paper researches a new vision of the telecommunication network of the future, its effects on human life and society, and the targets to achieve a new generation network (NwGN). In the paper, we also propose orientation towards an NwGN from the current networks, especially with Vietnam's telecommunications networks.

• 전자통신동향분석
• /
• 제35권4호
• /
• pp.34-41
• /
• 2020

The existing Von Neumann architecture places limits to data processing in AI, a booming technology. To address this issue, research is being conducted on computing architectures and artificial neural networks that simulate neurons and synapses, which are the hardware of the human brain. With high-speed, high-throughput data communication infrastructures, photonic solutions today are a mature industrial reality. In particular, due to the recent outstanding achievements of artificial neural networks, there is considerable interest in improving their speed and energy efficiency by exploiting photonic-based neuromorphic hardware instead of electronic-based hardware. This paper covers recent photonic neuromorphic studies and a classification of existing solutions (categorized into multilayer perceptrons, convolutional neural networks, spiking neural networks, and reservoir computing).

• ETRI Journal
• /
• 제33권3호
• /
• pp.299-309
• /
• 2011

Energy consumption is a key issue in body sensor networks (BSNs) since energy-constrained sensors monitor the vital signs of human beings in healthcare applications. In this paper, packet size optimization for BSNs has been analyzed to improve the efficiency of energy consumption. Existing studies on packet size optimization in wireless sensor networks cannot be applied to BSNs because the different operational characteristics of nodes and the channel effects of in-body and on-body propagation cannot be captured. In this paper, automatic repeat request (ARQ), forward error correction (FEC) block codes, and FEC convolutional codes have been analyzed regarding their energy efficiency. The hop-length extension technique has been applied to improve this metric with FEC block codes. The theoretical analysis and the numerical evaluations reveal that exploiting FEC schemes improves the energy efficiency, increases the optimal payload packet size, and extends the hop length for all scenarios for in-body and on-body propagation.