• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.6
• /
• pp.277-283
• /
• 2023

Vital signals provide essential information regarding the health status of individuals, thereby contributing to health management and medical research. Present monitoring methods, such as ECGs (Electrocardiograms) and smartwatches, demand proximity and fixed postures, which limit their applicability. To address this, Non-contact vital signal measurement methods, such as CW (Continuous-Wave) radar, have emerged as a solution. However, unwanted signal components and a stepwise processing approach lead to errors and limitations in heart rate detection. To overcome these issues, this study introduces an integrated neural network approach that combines noise removal, demodulation, and dominant-frequency detection into a unified process. The neural network employed for signal processing in this research adopts a MLP (Multi-Layer Perceptron) architecture, which analyzes the in-phase and quadrature signals collected within a specified time window, using two distinct input layers. The training of the neural network utilizes CW radar signals and reference heart rates obtained from the ECG. In the experimental evaluation, networks trained on different datasets were compared, and their performance was assessed based on loss and frequency accuracy. The proposed methodology exhibits substantial potential for achieving precise vital signals through non-contact measurements, effectively mitigating the limitations of existing methodologies.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.11 no.3
• /
• pp.159-166
• /
• 2007

This paper evaluates the soundness of structural elements using Wavelet Packet Transform (WPT). WPT is applied to the response acceleration of a framed structure which is subjected to earthquake load to decompose the response acceleration, then the energy of each component is calculated. The first five largest components in energy magnitude among the decomposed components are selected as input to an ANN to identify the damage location and severity. Two nodes in output layer yield damaged element and damage severity respectively. This method successfully evaluates the amount of damage and its location in the structure.

• Korean Journal of Agricultural Science
• /
• v.50 no.3
• /
• pp.399-406
• /
• 2023

In this study, we propose a ripe tomato bunch localization method based on convolutional neural networks, to be applied in robotic harvesting systems. Tomato images were obtained from a smart greenhouse at the Rural Development Administration (RDA). The sample images for training were extracted based on tomato maturity and resized to 128 × 128 pixels for use in the classification model. The model was constructed based on four-layer convolutional neural networks, and the classes were determined based on stage of maturity, using a Softmax classifier. The localization of the ripe tomato bunch region was indicated on a class activation map. The class activation map could show the approximate location of the tomato bunch but tends to present a local part or a large part of the ripe tomato bunch region, which could lead to poor performance. Therefore, we suggest a recursive method to improve the performance of the model. The classification results indicated that the accuracy, precision, recall, and F1-score were 0.98, 0.87, 0.98, and 0.92, respectively. The localization performance was 0.52, estimated by the Intersection over Union (IoU), and through input recursion, the IoU was improved by 13%. Based on the results, the proposed localization of the ripe tomato bunch area can be incorporated in robotic harvesting systems to establish the optimal harvesting paths.

• 한국해양학회지
• /
• v.27 no.4
• /
• pp.303-310
• /
• 1992

Studies on the distribution of nitrogenous compounds, and respiratory oxygen consumption rate were carried out in Masan Bay, Korea where large amount of industrial and domestic wastewaters are discharged. In August 1986 the surface layer was significantly influenced by freshwater input. Below the seasonal pycnocline, an oxygen-deficient condition developed in a large area of Masan Bay. Concentrations of DIN, DON and PN were 735.6, 1261.8 and 48.5 umol/l at the head, and 79.1, 73.0 and 39.5 umol/l at the mouth of the inner Masan Bay, respectively. Phytoplankton carbon production was 2,695 mgC/m$^2$/day at the mouth of inner Masan Bay. Dissolved oxygen contents were lower than 1 ml/l from 3 m depth in inner Masan Bay and from 10 m depth in the outer Masan Bay. The high concentration of ammonium and phosphate in the lower layer suggests the active degradation of organic materials in the bottom waters and leaching from sediments. The ERS activity was 232.1 ul O$_2$/l/h in the surface waters of the innermost part of Masan Bay and respiratory oxygen consumption is likely to proceed at a rate of 442 ml O$_2$/m$^2$/day in the bottom waters of this bay. Nitrate removal rate was estimated to be 0.25 umol/l/day via denitrification in the bottom waters of the Masan Waterway. It is estimated from the ETS activity that, at the mouth of inner Masan Bay, 9.3-10.5% of carbon fixed in the upper layer was decomposed below the themocline.

• The Journal of the Acoustical Society of Korea
• /
• v.33 no.1
• /
• pp.1-9
• /
• 2014

Coastal water including estuaries has distinctive environmental characteristics where sediments are transported and deposited by flowing river water, providing an environment in which fluid mud layers can be formed. Acoustic method is mostly used to detect or monitor the fluid mud layer. However, since sound propagating in this layer suffers severe attenuation, it is important to estimate the accurate attenuation coefficient for various concentrations of fluid mud layer for the successful use of the acoustic method. In this paper, measurement results of attenuation coefficient for 3.5, 5, and 7.5 MHz ultrasounds were presented. The measurements were made in a small-size water tank in which suspended sediment samples with various sediment concentrations were formed using kaolinite powder. The results were compared to the model predictions obtained by attenuation coefficient model in which the mean grain size (called as Mass-median-diameter, D50) was used as input parameter. There were reasonable agreements between measured attenuation coefficients and model outputs predicted using the particle range of D50 ${\pm}20%$. The comparison results imply that although the suspended sediments consist of various-sized particles, sound attenuation might be greatly influenced by amount of particle with a size which has a larger attenuation than that of any particle in the suspended sediments for the frequency used.

• The Journal of the Acoustical Society of Korea
• /
• v.30 no.8
• /
• pp.446-460
• /
• 2011

The use of induction motors has been recently increasing with automation in aeronautical and automotive industries, and it playes a significant role. This has motivated that many researchers have studied on developing fault detection and classification systems of an induction motor in order to minimize economical damage caused by its fault. With this reason, this paper proposed feature vector extraction methods based on STE (short-time energy)+SVD (singular value decomposition) and DCT (discrete cosine transform)+SVD techniques to early detect and diagnose faults of induction motors, and classified faults of an induction motor into different types of them by using extracted features as inputs of BPNN (back propagation neural network) and multi-layer SVM (support vector machine). When BPNN and multi-lay SVM are used as classifiers for fault classification, there are many settings that affect classification performance: the number of input layers, the number of hidden layers and learning algorithms for BPNN, and standard deviation values of Gaussian radial basis function for multi-layer SVM. Therefore, this paper quantitatively simulated to find appropriate settings for those classifiers yielding higher classification performance than others.

• Applied Microscopy
• /
• v.34 no.4
• /
• pp.285-294
• /
• 2004

We investigated the morphology, distribution and synaptic connectivity of cholinergic neurons in the mouse retina by immunocytochemistry, using antisera against choline acetyltransferase (ChAT). ChAT-immunoreactive amacrine cells fall into two groups according to the localization of their somas in the retina: one is situated in the inner nuclear layer (INL), near the border of the inner plexiform layer (IPL), and the other is displaced in the ganglion cell layer (GCL). The dendrites of amacrine cells from the INL ramify in sublamina a and that of the displaced amacrine cells ramify in sublamina b of the IPL. Double labeling with an antisera against ChAT and r-aminobutyric acid (GABA) demonstrated that these labeled cells formed a subpopulation of GABAergic amacrine cells. The synaptic connectivity of ChAT-immunoreactive amacrine cells was identified in the IPL by electron microscopy. The most frequent synaptic input of ChAT-labeled amacrine cells was from bipolar cells in both sublaminae a and b of the IPL, followed by labeled amacrine cells and unlabeled amacrine cells. Their primary output targets were onto ganglion cells in both sublaminae a and b and output onto ganglion cells was more frequently observed in sublamina b of the IPL. Our results suggest that cholinergic amacrine cells in the mouse retina are very similar to their counter parts in other mammals, and they can attribute a major role in the pathway feeding into directionally selective ganglion cells.

• Applied Chemistry for Engineering
• /
• v.25 no.2
• /
• pp.174-180
• /
• 2014

A plasma hydrophobic coating on commercial fish feed was conducted to prolong the floating time of feed, thereby enhancing the feed consumption rate and reducing the contamination of water in fish farms. The hydrophobic coating on the fish feed was prepared using an atmospheric-pressure dielectric barrier discharge (DBD) plasma with hexamethyldisiloxane (HMDSO), toluene and n-hexane as the precursors. The effect of the parameters such as input power, precursor type and coating time on the coating performance were examined. The physicochemical properties of the coating layer were analyzed using a Fourier transform infrared (FTIR) spectrometer and a contact angle (CA) analyzer. The water CA increased after the coating preparation, indicating that the surface changed from hydrophilic to hydrophobic. The FTIR characterization revealed that the hydrophobic layer was comprised of functional groups such as $CH_3$, Si-O-Si and Si-C. As a result of the hydrophobic coating, the floating time of the fish feed increased from several seconds to 3 minutes, which suggested that the plasma coating method could be a viable means for practical applications. Compared to the water CA measured as soon as the coating layer was prepared, the 6-day aged sample exhibited a substantial CA increase, confirming the aging effect on the improvement of the hydrophobicity.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.7
• /
• pp.970-976
• /
• 2007

In this study, Polynomial Network Pattern Classifier(PNC) based on Fuzzy Inference Mechanism is designed and its parameters such as learning rate, momentum coefficient and fuzzification coefficient are optimized by means of Particle Swarm Optimization. The proposed PNC employes a partition function created by Fuzzy C-means(FCM) clustering as an activation function in hidden layer and polynomials weights between hidden layer and output layer. Using polynomials weights can help to improve the characteristic of the linear classification of basic neural networks classifier. In the viewpoint of linguistic analysis, the proposed classifier is expressed as a collection of "If-then" fuzzy rules. Namely, architecture of networks is constructed by three functional modules that are condition part, conclusion part and inference part. The condition part relates to the partition function of input space using FCM clustering. In the conclusion part, a polynomial function caries out the presentation of a partitioned local space. Lastly, the output of networks is gotten by fuzzy inference in the inference part. The proposed PNC generates a nonlinear discernment function in the output space and has the better performance of pattern classification as a classifier, because of the characteristic of polynomial based fuzzy inference of PNC.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.26 no.3
• /
• pp.563-572
• /
• 2016

Recently, a plenty of credit card payment protocols have been proposed in Korea. Several features of proposed protocols include: using passwords for user authentication in stead of official certificate for authenticity, and no need to download additional security module via ActiveX into user's devices. In this paper, we suggest two new credit card payment protocols that use both SSL(Security Socket Layer) as a standardized secure transaction protocol and password authentication to perform online shopping and payment. The first one is for the case where online shopping mall is different from PG(Payment Gateway) and can be compared to PayPal-based payment methods, and the second one is for the case where online shopping mall is the same as PG and thus can be compared to Amazon-like methods. Two proposed protocols do not require users to perform any pre-registration process which is separate from an underlying shopping process, instead users can perform both shopping and payment into a single process in a convenient way. Also, users are asked to input a distinct payment password, which increases the level of security in the payment protocols. We believe that two proposed protocols can help readers to better understand the recent payment protocols that are suggested by various vendors, and to analyze the security of their payment protocols.