• Journal of the Korea Society of Computer and Information
• /
• v.29 no.2
• /
• pp.31-41
• /
• 2024

Deep learning models show excellent performance in tasks such as image classification and object detection in the field of computer vision, and are used in various ways in actual industrial sites. Recently, research on improving robustness has been actively conducted, along with pointing out that this deep learning model is vulnerable to hostile examples. A hostile example is an image in which small noise is added to induce misclassification, and can pose a significant threat when applying a deep learning model to a real environment. In this paper, we tried to confirm the robustness of the edge-learning classification model and the performance of the adversarial example detection model using it for adversarial examples of various algorithms. As a result of robustness experiments, the basic classification model showed about 17% accuracy for the FGSM algorithm, while the edge-learning models maintained accuracy in the 60-70% range, and the basic classification model showed accuracy in the 0-1% range for the PGD/DeepFool/CW algorithm, while the edge-learning models maintained accuracy in 80-90%. As a result of the adversarial example detection experiment, a high detection rate of 91-95% was confirmed for all algorithms of FGSM/PGD/DeepFool/CW. By presenting the possibility of defending against various hostile algorithms through this study, it is expected to improve the safety and reliability of deep learning models in various industries using computer vision.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.3
• /
• pp.44-51
• /
• 2009

This paper addresses a face verification method using the nonlinear composite filter. This face verification process can be simple and speedy because it does not require any reprocessing such as face detection, alignment or cropping. The optimum nonlinear composite filter is derived by minimizing the output energy due to additive noise and an input scene while maintaining the outputs of training images constant. The filter is equipped with the discrimination capability and the robustness to additive noise by minimizing the outputs of the input scene and the noise, respectively. We build the nonlinear composite filter with two training images and compare the filter with the conventional synthetic discriminant function (SDF) filter. The receiver operating characteristics (ROC) curves are presented as a metric for the performance evaluation. According to the experimental results the optimum nonlinear composite filter is shown to be a robust scheme for face verification in low resolution and noise environments.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.4
• /
• pp.2124-2148
• /
• 2017

Accurate facial expression recognition (FER) requires reliable signal filtering and the effective feature extraction. Considering these requirements, this paper presents a novel approach for FER which is robust to noise. The main contributions of this work are: First, to preserve texture details in facial expression images and remove image noise, we improved the anisotropic diffusion filter by adjusting the diffusion coefficient according to two factors, namely, the gray value difference between the object and the background and the gradient magnitude of object. The improved filter can effectively distinguish facial muscle deformation and facial noise in face images. Second, to further improve robustness, we propose a new feature descriptor based on a combination of the Histogram of Oriented Gradients with the Canny operator (Canny-HOG) which can represent the precise deformation of eyes, eyebrows and lips for FER. Third, Canny-HOG's block and cell sizes are adjusted to reduce feature dimensionality and make the classifier less prone to overfitting. Our method was tested on images from the JAFFE and CK databases. Experimental results in L-O-Sam-O and L-O-Sub-O modes demonstrated the effectiveness of the proposed method. Meanwhile, the recognition rate of this method is not significantly affected in the presence of Gaussian noise and salt-and-pepper noise conditions.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.4
• /
• pp.469-474
• /
• 2011

This paper presents the intelligent tracking method for maneuvering target using the positional error compensation of the maneuvering target. The difference between measured point and predict point is separated into acceleration and noise. K-means clustering and TS fuzzy system are used to get the optimal acceleration value. The membership function is determined for acceleration and noise which are divided by K-means clustering and the characteristics of the maneuvering target is figured out. Divided acceleration and noise are used in the tracking algorithm to compensate computational error. While calculating expected value, the non-linearity of the maneuvering target is recognized as linear one by dividing acceleration and the capability of Kalman filter is kept in the filtering process. The error for the non-linearity is compensated by approximated acceleration. The proposed system improves the adaptiveness and the robustness by adjusting the parameters in the membership function of fuzzy system. Procedures of the proposed algorithm can be implemented as an on-line system. Finally, some examples are provided to show the effectiveness of the proposed algorithm.

• Journal of the Institute of Electronics Engineers of Korea TE
• /
• v.39 no.2
• /
• pp.44-50
• /
• 2002

Conventional acoustic echo cancelers using ES(Exponentially weighted Stepsize) algorithm have simple operational configuration and fast convergence speed batter then NLMS algorithm, but they are very weak in external noise because ES algorithm updates filter taps using an average energy reduction rate of room impulse response in specific acoustical condition. So, a new configuration of acoustic echo canceler with stepsize generator and selector is proposed in this thesis. The proposed stepsize generator and selector improve conventional acoustic echo canceler's weakness in external noise and improve the system robustness. The stepsize generator generates additional stepsize value using moving averager, which is the residual noise energy of error signal multiplied by constant ${\gamma}$. The stepsize selector selects the stepsize value that has better performance in an acoustic echo canceler using a coefficient decision factor ${\Delta}_{differ}$ The simulation results show that the proposed algorithm reduces residual error by 5[dB] to 10[dB], improves misadjustment regardless of external noise's SNR. 

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.15 no.1
• /
• pp.27-35
• /
• 2022

In this paper we propose white noise adding method to prevent missclassification of deep learning system by adversarial attacks. The proposed method is that adding white noise to input image that is benign or adversarial example. The experimental results are showing that the proposed method is robustness to 3 adversarial attacks such as FGSM attack, BIN attack and CW attack. The recognition accuracies of Resnet model with 18, 34, 50 and 101 layers are enhanced when white noise is added to test data set while it does not affect to classification of benign test dataset. The proposed model is applicable to defense to adversarial attacks and replace to time- consuming and high expensive defense method against adversarial attacks such as adversarial training method and deep learning replacing method.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.3
• /
• pp.155-163
• /
• 2023

Real-time monitoring technology is critical for ensuring the safety and reliability of nuclear power plant structures. However, the current seismic monitoring system has limited system identification capabilities such as modal parameter estimation. To obtain global behavior data and dynamic characteristics, multiple sensors must be optimally placed. Although several studies on optimal sensor placement have been conducted, they have primarily focused on civil and mechanical structures. Nuclear power plant structures require robust signals, even at low signal-to-noise ratios, and the robustness of each mode must be assessed separately. This is because the mode contributions of nuclear power plant containment buildings are concentrated in low-order modes. Therefore, this study proposes an optimal sensor placement methodology that can evaluate robustness against noise and the effects of each mode. Indicators, such as auto modal assurance criterion (MAC), cross MAC, and mode shape distribution by node were analyzed, and the suitability of the methodology was verified through numerical analysis.

• Journal of Korea Multimedia Society
• /
• v.6 no.7
• /
• pp.1149-1158
• /
• 2003

Content adaptive watermark embedding algorithm using a stochastic image model in the multiwavelet transform is proposed in this paper. A watermark is embedded into the perceptually significant coefficients (PSCs) of each subband using multiwavelet transform. The PSCs in high frequency subband are selected by SSQ, that is, by setting the thresholds as the one half of the largest coefficient in each subband. The perceptual model is applied with a stochastic approach based on noise visibility function (NVF) that has local image properties for watermark embedding. This model uses stationary Generalized Gaussian model characteristic because watermark has noise properties. The watermark estimation use shape parameter and variance of subband region. it is derive content adaptive criteria according to edge and texture, and flat region. The experiment results of the proposed watermark embedding method based on multiwavelet transform techniques were found to be excellent invisibility and robustness.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.3
• /
• pp.91-98
• /
• 2018

Automatic detection of pig wasting diseases is an important issue in the management of group-housed pigs. In particular, porcine respiratory diseases are one of the main causes of mortality among pigs and loss of productivity in intensive pig farming. In this paper, we propose a noise-robust system for the early detection and recognition of pig wasting diseases using sound data. In this method, first we convert one-dimensional sound signals to two-dimensional gray-level images by normalization, and extract texture images by means of dominant neighborhood structure technique. Lastly, the texture features are then used as inputs of convolutional neural networks as an early anomaly detector and a respiratory disease classifier. Our experimental results show that this new method can be used to detect pig wasting diseases both economically (low-cost sound sensor) and accurately (over 96% accuracy) even under noise-environmental conditions, either as a standalone solution or to complement known methods to obtain a more accurate solution.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.4
• /
• pp.15-22
• /
• 2011

Principal component analysis (PCA) is a well-known method for dimensionality reduction and feature extraction while maintaining most of the variation in data. Although PCA has been applied in many areas successfully, it is sensitive to outliers and only valid for Gaussian distributions. Several variants of PCA have been proposed to resolve noise sensitivity and, among the variants, improved robust fuzzy PCA (RF-PCA2) demonstrated promising results. RF-PCA, however, is still a linear algorithm that cannot accommodate non-Gaussian distributions. In this paper, a non-linear algorithm that combines RF-PCA2 and kernel PCA (K-PCA), called improved robust kernel fuzzy PCA (RKF-PCA2), is introduced. The kernel methods make it to accommodate non-Gaussian distributions. RKF-PCA2 inherits noise robustness from RF-PCA2 and non-linearity from K-PCA. RKF-PCA2 outperforms previous methods in handling non-Gaussian distributions in a noise robust way. Experimental results also support this.