• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.2
• /
• pp.149-162
• /
• 2005

This paper suggest a way to detect a specific wanted figure in public places such as subway stations and banks by comparing color face images extracted from the real time CCTV with the face images of designated specific figures. Assuming that the characteristic of the surveillance camera allows the face information in screens to change arbitrarily and to contain information on numerous faces, the accurate detection of the face area was focused. To solve this problem, the normalization work using subsampling with $20{\times}20$ pixels on arbitrary face images, which is based on the Perceptron Neural Network model suggested by R. Rosenblatt, created the effect of recogning the whole face. The optimal linear filter and the histogram shaper technique were employed to minimize the outside interference such as lightings and light. The addition operation of the egg-shaped masks was added to the pre-treatment process to minimize unnecessary work. The images finished with the pre-treatment process were divided into three reception fields and the information on the specific location of eyes, nose, and mouths was determined through the neural network. Furthermore, the precision of results was improved by constructing the three single-set network system with different initial values in a row.

• Smart Media Journal
• /
• v.11 no.1
• /
• pp.58-66
• /
• 2022

As Internet technology develops, SNS users are increasing. As SNS becomes popular, SNS-type crimes using the influence and anonymity of social networks are increasing day by day. In this paper, we propose a fake account classification method that applies machine learning and deep learning to statistical and image data for fake accounts classification. SNS account data used for training was collected by itself, and the collected data is based on statistical data and image data. In the case of statistical data, machine learning and multi-layer perceptron were employed to train. Furthermore in the case of image data, a convolutional neural network (CNN) was utilized. Accordingly, it was confirmed that the overall performance of account classification was significantly meaningful.

• Journal of the Korea Concrete Institute
• /
• v.27 no.4
• /
• pp.331-341
• /
• 2015

The main purpose of this study is reducing the cost and effort for characterization of tensile strength of fiber reinforced concrete, in order to use in structural design. For this purpose, in this study, test for fiber reinforced concrete was carried out. Because fiber reinforced concrete is consisted of diverse material, it is hard to define the correlation between mix proportions and strength. Therefore, compressive strength test and tensile strength test were carried out for the range of smaller than 100 MPa of compressive strength and 0.25~1% of steel fiber volume fraction. as a results of test, two types of tensile strength were highly affected by compressive strength of concrete. However, increase rate of tensile strength was decreased with increase of compressive strength. Increase rate of tensile strength was decreased with increase of fiber volume fraction. Database was constructed using previous research data. Because estimation equations for tensile strength of fiber reinforced concrete should be multiple variable function, linear regression is hard to apply. Therefore, in this study, we decided to use the ANN(Artificial Neural Network). ANN was constructed using multiple layer perceptron architecture. Sigmoid function was used as transfer function and back propagation training method was used. As a results of prediction using artificial neural network, predicted values of test data and previous research which was randomly selected were well agreed with each other. And the main effective parameters are water-cement ratio and fiber volume fraction.

• Journal of the Institute of Convergence Signal Processing
• /
• v.20 no.2
• /
• pp.70-77
• /
• 2019

Breast ultrasound readings are very important to diagnose early breast cancer. In Ultrasonic inspection, it shows a significant difference in image quality depending on the ultrasonic equipment, and there is a large difference in diagnosis depending on the experience and skill of the inspector. Therefore, objective criteria are needed for accurate diagnosis and treatment. In this study, we analyzed texture characteristics by applying GLCM (Gray Level Co-occurrence Matrix) algorithm and extracted characteristic parameters and diagnosed breast cancer using neural network classifier. Breast ultrasound images were classified into normal, benign and malignant tumors and six texture parameters were extracted. Fourteen cases of normal, malignant and benign tumor diagnosed by mammography were studied by using the extracted six parameters and learning by multi - layer perceptron neural network back propagation learning method. As a result of classification using 51 normal images, 62 benign tumor images, and 74 malignant tumor images of the learned model, the classification rate was 95.2%.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.48 no.6
• /
• pp.1-8
• /
• 2011

In multimedia fingerprinting field, it is many used a code based on BIBD, which has a strong resiliency of anti-collusion. When a collusion-attack code is generated with a logical XOR operation using the code based on BIBD, then some cases are occurred that a colluded code could be generated to the same fingerprint of non-colluder on the other hand, the colluder is decided to the non-colluder so that he would be excepted in the colluder tracing. For solving the serious problem of the wrong decision of the colluder tracing in this paper, XOR-ACC is implemented using multi-layer perceptron neural network among (AND, OR, XOR and Averaging)-ACC by the measured correlation coefficient. Through the experiment, it confirms that XOR-ACC efficiency of multimedia fingerprinting code{7,3,1} based on BIBD is improved to 88.24% from the conventional 41.18%, so that a ratio of the colluder tracing is also improved to 100% from the conventional 53%. As a result, it could be traced and decided completely a sectional colluder and non-colluder about the collusion attacks.

• Journal of KIISE:Software and Applications
• /
• v.34 no.10
• /
• pp.910-917
• /
• 2007

Of the possible interactions between human and robot, touch is an important means of providing human beings with emotional relief. However, most previous studies have focused on interactions based on voice and images. In this paper. a method of recognizing human touching behaviors is proposed for developing a robot that can naturally interact with humans through touch. In this method, the recognition process is divided into pre-process and recognition Phases. In the Pre-Process Phase, recognizable characteristics are calculated from the data generated by the touch detector which was fabricated using force sensors. The force sensor used an FSR (force sensing register). The recognition phase classifies human touching behaviors using a multi-layer perceptron which is a neural network model. Experimental data was generated by six men employing three types of human touching behaviors including 'hitting', 'stroking' and 'tickling'. As the experimental result of a recognizer being generated for each user and being evaluated as cross-validation, the average recognition rate was 82.9% while the result of a single recognizer for all users showed a 74.5% average recognition rate.

• Journal of the Institute of Convergence Signal Processing
• /
• v.13 no.3
• /
• pp.136-141
• /
• 2012

One of the main reasons for serious road accidents is driving while drowsy. For this reason, drowsiness detection and warning system for drivers has recently become a very important issue. Monitoring physiological signals provides the possibility of detecting features of drowsiness and fatigue of drivers. One of the effective signals is to measure electroencephalogram (EEG) signals and electrooculogram (EOG) signals. The aim of this study is to extract drowsiness-related features from a set of EEG signals and to classify the features into three states: alertness, drowsiness, sleepiness. This paper proposes a neural-network-based drowsiness detection system using Linear Predictive Coding (LPC) coefficients as feature vectors and Multi-Layer Perceptron (MLP) as a classifier. Samples of EEG data from each predefined state were used to train the MLP program by using the proposed feature extraction algorithms. The trained MLP program was tested on unclassified EEG data and subsequently reviewed according to manual classification. The classification rate of the proposed system is over 96.5% for only very small number of samples (250ms, 64 samples). Therefore, it can be applied to real driving incident situation that can occur for a split second.

• Journal of KIISE:Software and Applications
• /
• v.31 no.4
• /
• pp.457-466
• /
• 2004

Recommendation is to offer information which fits user's interests and tastes to provide better services and to reduce information overload. It recently draws attention upon Internet users and information providers. The collaborative filtering is one of the widely used methods for recommendation. It recommends an item to a user based on the reference users' preferences for the target item or the target user's preferences for the reference items. In this paper, we propose a neural network based collaborative filtering method. Our method builds a model by learning correlation between users or items using a multi-layer perceptron. We also investigate integration of diverse information to solve the sparsity problem and selecting the reference users or items based on similarity to improve performance. We finally demonstrate that our method outperforms the existing methods through experiments using the EachMovie data.

• Steel and Composite Structures
• /
• v.45 no.2
• /
• pp.205-218
• /
• 2022

Proper calculation of splitting tensile strength (STS) of concrete has been a crucial task, due to the wide use of concrete in the construction sector. Following many recent studies that have proposed various predictive models for this aim, this study suggests and tests the functionality of three hybrid models in predicting the STS from the characteristics of the mixture components including cement compressive strength, cement tensile strength, curing age, the maximum size of the crushed stone, stone powder content, sand fine modulus, water to binder ratio, and the ratio of sand. A multi-layer perceptron (MLP) neural network incorporates invasive weed optimization (IWO), cuttlefish optimization algorithm (CFOA), and electrostatic discharge algorithm (ESDA) which are among the newest optimization techniques. A dataset from the earlier literature is used for exploring and extrapolating the STS behavior. The results acquired from several accuracy criteria demonstrated a nice learning capability for all three hybrid models viz. IWO-MLP, CFOA-MLP, and ESDA-MLP. Also in the prediction phase, the prediction products were in a promising agreement (above 88%) with experimental results. However, a comparative look revealed the ESDA-MLP as the most accurate predictor. Considering mean absolute percentage error (MAPE) index, the error of ESDA-MLP was 9.05%, while the corresponding value for IWO-MLP and CFOA-MLP was 9.17 and 13.97%, respectively. Since the combination of MLP and ESDA can be an effective tool for optimizing the concrete mixture toward a desirable STS, the last part of this study is dedicated to extracting a predictive formula from this model.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.30 no.6
• /
• pp.1291-1300
• /
• 2020

Recently, a deep learning-based non-profiling side-channel analysis was proposed. The deep learning-based non-profiling analysis is a technique that trains a neural network model for all guessed keys and then finds the correct secret key through the difference in the training metrics. As the performance of non-profiling analysis varies greatly depending on the neural network training model design, a correct model design criterion is required. This paper describes the two types of loss functions and eight labeling methods used in the training model design. It predicts the analysis performance of each labeling method in terms of non-profiling analysis and power consumption model. Considering the characteristics of non-profiling analysis and the HW (Hamming Weight) power consumption model is assumed, we predict that the learning model applying the HW label without One-hot encoding and the Correlation Optimization (CO) loss will have the best analysis performance. And we performed actual analysis on three data sets that are Subbytes operation part of AES-128 1 round. We verified our prediction by non-profiling analyzing two data sets with a total 16 of MLP-based model, which we describe.