• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.17 no.2
• /
• pp.113-120
• /
• 2024

The population sparseness index(PSI) is being utilized to describe the functioning of internal layers in artificial neural networks from the perspective of neurons, shedding light on the black-box nature of the network's internal operations. There is research indicating a positive correlation between the PSI and performance in each layer of convolutional neural network models for image classification. In this study, we observed the internal operations of a convolutional neural network when adversarial examples were applied. The results of the experiments revealed a similar pattern of positive correlation for adversarial examples, which were modified to maintain 5% accuracy compared to applying benign data. Thus, while there may be differences in each adversarial attack, the observed PSI for adversarial examples demonstrated consistent positive correlations with benign data across layers.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.1
• /
• pp.1-7
• /
• 2023

Convolutional neural networks have already been applied to various fields beyond human visual processing capabilities in the image processing area. However, they are exposed to a severe risk of deteriorating model performance due to the appearance of adversarial attacks. In addition, defense technology to respond to adversarial attacks is effective against the attack but is vulnerable to other types of attacks. Therefore, to respond to an adversarial attack, it is necessary to analyze how the performance of the adversarial attack deteriorates through the process inside the convolutional neural network. In this study, the adversarial attack of the Alexnet and VGG11 models was analyzed using the population sparseness index, a measure of neuronal activity in neurophysiology. Through the research, it was observed in each layer that the population sparsity index for adversarial examples showed differences from that of benign examples.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.8
• /
• pp.936-942
• /
• 2019

Autonomous driving of drone indoor must move along a narrow path and overcome other factors such as lighting, topographic characteristics, obstacles. In addition, it is difficult to operate the drone in the hallway because of insufficient texture and the lack of its diversity comparing with the complicated environment. In this paper, we study an autonomous drone navigation using Convolution Neural Network(CNN) in indoor environment. The proposed method receives an image from the front camera of the drone and then steers the drone by predicting the next path based on the image. As a result of a total of 38 autonomous drone navigation tests, it was confirmed that a drone was successfully navigating in the indoor environment by the proposed method without hitting the walls or doors in the hallway.

• Journal of KIISE
• /
• v.44 no.7
• /
• pp.668-673
• /
• 2017

Deep neural networks have shown remarkable performance in various fields of pattern recognition such as voice recognition, image recognition and object detection. However, underlying mechanisms of the network have not been fully revealed. In this paper, we focused on empirical analysis of the network parameters. The Faster R-CNN(region-based convolutional neural network) was used as a baseline network of our work and three important parameters were analyzed: the dropout ratio which prevents the overfitting of the neural network, the size of the anchor boxes and the activation function. We also compared the performance of dropout and batch normalization. The network performed favorably when the dropout ratio was 0.3 and the size of the anchor box had not shown notable relation to the performance of the network. The result showed that batch normalization can't entirely substitute the dropout method. The used leaky ReLU(rectified linear unit) with a negative domain slope of 0.02 showed comparably good performance.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.06a
• /
• pp.163-165
• /
• 2017

본 논문에서는 카메라 잡음 제거에 딥 러닝 알고리즘을 적용하는 연구를 진행하였다. 합성된 가우시언 잡음에 대하여 좋은 잡음 제거 성능을 보이는 DnCNN(Denoising Convolutional Network)를 이용하여 카메라 잡음을 제거하는 학습과 실험을 진행하였으며, 기준 실험으로는 RGB 색공간의 3채널 모두에 대하여 학습한 신경망(Neural Network)을 사용하였고, 본 논문의 실험에서는 그레이 이미지에 대하여 학습한 신경망을 사용하였다. 신경망의 평가를 위하여 딥 러닝 알고리즘 입력 이미지를 RGB 색공간(RGB Color Space)과 YCbCr 색공간(YCbCr Color Space) 2가지 색공간으로 표현하여 사용하였고, 입력 이미지에 노이즈를 첨가하기 위해 가우시안 노이즈(Gaussian Noise)를 이용하였다. 또한 가우시안 잡음과 다른 성질을 갖는 실제 카메라 잡음에 대해서도 학습과 테스트를 진행하였다.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.3
• /
• pp.449-458
• /
• 2023

Many people think that sign language is only for people who are deaf and cannot speak, but of course it is necessary for people who want to talk with them. One of the biggest challenges in ASL(American Sign Language) alphabet recognition is the high inter-class similarities and high intra-class variance. In this paper, we proposed an architecture that can overcome these two problems, which performs similarity learning to reduces inter-class similarities and intra-class variance between images. The proposed architecture consists of the same convolutional neural network with a double configuration that shares parameters (weights and biases) and also applies the Keras API to reduce similarity learning and variance through this pathway. The similarity learning results the use of the dual CNN shows that the accuracy is improved by reducing the similarity and variability between classes by not including the poor results of the two classes.

• Journal of Korea Society of Industrial Information Systems
• /
• v.22 no.5
• /
• pp.17-24
• /
• 2017

Recently Deep-learning, Technologies using Large or Deep Artificial Neural Networks, have Shown Remarkable Performance, and the Increasing Size of the Network Contributes to its Performance Improvement. However, the Increase in the Size of the Neural Network Leads to an Increase in the Calculation Amount, which Causes Problems Such as Circuit Complexity, Price, Heat Generation, and Real-time Restriction. In This Paper, We Propose and Test a Method to Reduce the Number of Network Connections by Effectively Pruning the Redundancy in the Connection and Showing the Difference between the Performance and the Desired Range of the Original Neural Network. In Particular, we Proposed a Simple Method to Improve the Performance by Re-learning and to Guarantee the Desired Performance by Allocating the Error Rate per Layer in Order to Consider the Difference of each Layer. Experiments have been Performed on a Typical Neural Network Structure such as FCN (full connection network) and CNN (convolution neural network) Structure and Confirmed that the Performance Similar to that of the Original Neural Network can be Obtained by Only about 1/10 Connection.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.153-155
• /
• 2019

본 논문에서는 컨볼루션 신경망을 이용하여 예측 된 공간 오디오 큐를 이용한 오디오 채널 확장 기술을 소개한다. 오디오 채널 확장 기술은 일반적인 스테레오 신호에 적용되어 5.1 레이아웃과 같은 고차원 오디오 신호를 생성하는 기술이다. 스테레오 신호에서 채널을 확장하기 위해 스테레오 신호에서 공간 큐를 예측하고 예측 공간 큐의 방향에 따라 5.1 채널 신호의 스펙트럼 구성 요소를 할당하여 다중 채널 신호를 합성한다. 제안된 방식으로 생성된 5.1 채널 신호는 원 5.1 채널과 유사한 공간 정보 합성 능력과 스테레오 대비 주관적 선호도가 개선된 음질을 제공한다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.73-75
• /
• 2019

인터레이스 영상은 지난 수 십 년간 방송 및 비디오 레코딩 등에 광범위하게 사용되고 있으며 디인터레이스의 성능을 향상 시키기 위한 많은 연구가 이루어졌다. 이를 위한 것으로써 본 논문에서는 컨볼루션 신경망을 이용한 비디오 디인터레이스 기법을 제안한다. 제안한 방법은 SKIP 연결을 사용하여 낮은 수준 특징 정보를 뒷 단의 레이어까지 전달함으로써 성능 향상을 달성하였다. 실험 결과는 FFMPEG 에서 제공하는 디인터레이스 기법에 비해 전 영상에 걸쳐 우수한 성능을 제공하며, 특히 복잡한 영상에서 기존 알고리즘 대비 큰 폭의 성능향상을 보인다.

• Journal of the Korean Society of Mechanical Technology
• /
• v.20 no.6
• /
• pp.776-782
• /
• 2018

After CNN basic structure was introduced by LeCun in 1989, there has not been a major structure change except for more deep network until recently. The deep network enhances the expression power due to improve the abstraction ability of the network, and can learn complex problems by increasing non linearity. However, the learning of a deep network means that it has vanishing gradient or longer learning time. In this study, we proposes a CNN structure with MLP layer. The proposed CNNs are superior to the general CNN in their classification performance. It is confirmed that classification accuracy is high due to include MLP layer which improves non linearity by experiment. In order to increase the performance without making a deep network, it is confirmed that the performance is improved by increasing the non linearity of the network.