• 터널과지하공간
• /
• 제30권5호
• /
• pp.462-472
• /
• 2020

암종 분류은 현장의 지질학적 또는 지반공학적 특성 파악을 위해 요구되는 매우 기본적인 행위이나 암석의 성인, 지역, 지질학적 이력 특성에 따라 동일 암종이라 하여도 매우 다양한 형태와 색 조성을 보이므로 깊은 지질학적 학식과 경험 없이는 쉬운 일은 아니다. 또한, 다른 여러 분야의 분류 작업에서 딥러닝 영상 처리 기법들이 성공적으로 적용되고 있으며, 지질학적 분류나 평가 분야에서도 딥러닝 기법의 적용에 대한 관심이 증대되고 있다. 따라서, 본 연구에서는 동일 암종임에도 다양한 형태와 색을 갖게 되는 실제 상황을 감안하여, 정확한 자동 암종 분류를 위한 딥러닝 기법의 적용 가능성에 대해 검토하였다. 이러한 기법은 향후에 현장 암종분류 작업을 수행하는 현장 기술자들을 지원할 수 있는 효과적인 툴로 활용 가능할 것이다. 본 연구에서 사용된 딥러닝 알고리즘은 매우 깊은 네트워크 구조로 객체 인식과 분류를 할 수 있는 것으로 잘 알려진 'ResNet' 계열의 딥러닝 알고리즘을 사용하였다. 적용된 딥러닝에서는 10개의 암종에 대한 다양한 암석 이미지들을 학습시켰으며, 학습 시키지 않은 암석 이미지들에 대하여 84% 수준 이상의 암종 분류 정확도를 보였다. 본 결과로 부터 다양한 성인과 지질학적 이력을 갖는 다양한 형태와 색의 암석들도 지질 전문가 수준으로 분류해 낼 수 있는 것으로 파악되었다. 나아가 다양한 지역과 현장에서 수집된 암석의 이미지와 지질학자들의 분류 결과가 학습데이터로 지속적으로 누적이 되어 재학습에 반영된다면 암종분류 성능은 자동으로 향상될 것이다.

• 방송공학회논문지
• /
• 제26권4호
• /
• pp.409-418
• /
• 2021

컨포멀 코팅은 PCB(Printed Circuit Board)를 보호하는 기술로 PCB의 고장을 최소화한다. 코팅의 결함은 PCB의 고장과 연결되기 때문에 성공적인 컨포멀 코팅 조건을 만족하기 위해서 코팅면에 기포가 발생했는지 검사한다. 본 논문에서는 영상 신호 처리를 적용하여 고위험군의 문제성 기포를 검출하는 알고리즘을 제안한다. 알고리즘은 문제성 기포의 후보를 구하는 단계와 후보를 검증하는 단계로 구성된다. 기포는 가시광 영상에서 나타나지 않지만, UV(Ultra Violet) 광원에서는 육안으로 구별이 가능하다. 특히, 문제성 기포의 중심은 밝기가 어둡고 테두리는 높은 밝기를 가진다. 이러한 밝기 특성을 논문에서는 협곡과 산맥 특징이라 부르고 두 가지 특징이 동시에 나타나는 영역을 문제성 기포의 후보라 하였다. 그러나 후보 중에는 기포가 아닌 후보가 존재할 수 있기 때문에 후보를 검증하는 단계가 필요하다. 후보 검증 단계에서는 합성곱 신경망 모델을 이용하였고, ResNet이 다른 모델과 비교하였을 때 성능이 가장 우수하였다. 본 논문에서 제시한 알고리즘은 정확률(Precision) 0.805, 재현율(Recall) 0.763, F1-점수(F1-score) 0.767의 성능을 보였고, 이러한 결과는 기포 검사 자동화에 대한 충분한 가능성을 보여준다.

• 정보보호학회논문지
• /
• 제28권4호
• /
• pp.847-857
• /
• 2018

악성코드의 수가 가파르게 증가하면서 기업 및 공공기관, 금융기관, 병 의원 등을 타깃으로 한 사이버 공격 피해 사례가 늘어나고 있다. 이러한 흐름에 따라 학계와 보안 업계에서는 악성코드 탐지를 위한 다양한 연구를 진행하고 있다. 최근 들어서는 딥러닝을 비롯해 머신러닝 기법을 적용하는 형태의 연구가 많이 진행되는 추세다. 이 중 합성곱 신경망(CNN: Convolutional Neural Network), ResNet 등을 이용한 악성코드 분류 연구의 경우에는 기존의 분류 방법에 비해 정확도가 크게 향상된 것을 확인할 수 있다. 그러나 타깃 공격의 특징 중 하나는 사용된 악성코드가 불특정 다수를 상대로 광범위하게 퍼뜨리는 형태가 아닌, 특정 대상을 타깃으로 한 맞춤형 악성코드라는 점이다. 이러한 유형의 악성코드는 그 수가 많지 않기 때문에 기존에 연구되어온 머신러닝이나 딥러닝 기법을 적용하기에 한계가 있다. 본 논문은 타깃형 악성코드와 같이 샘플의 양이 부족한 상황에서 악성코드를 분류하는 방법에 대해 다루고 있다. 메모리가 추가된 신경망(MANN: Memory Augmented Neural Networks) 모델을 이용하였고 각 그룹별 20개의 소량 데이터로 구성되어 있는 악성코드 데이터셋에 대해 최대 97%까지 정확도로 분류할 수 있음을 확인하였다.

• 대한조선학회논문집
• /
• 제58권3호
• /
• pp.129-136
• /
• 2021

Cavitation erosion is one of the major factors causing damage by lowering the structural strength of the marine propeller and the risk of it has been qualitatively evaluated by each institution with their own criteria based on the experiences. In this study, in order to quantitatively evaluate the risk of cavitation erosion on the propeller, we implement a deep learning algorithm based on a convolutional neural network. We train and verify it using the model tests results, including cavitation characteristics of various ship types. Here, we adopt the validated well-known networks such as VGG, GoogLeNet, and ResNet, and the results are compared with the expert's qualitative prediction results to confirm the feasibility of the prediction algorithm using a convolutional neural network.

• Journal of Multimedia Information System
• /
• 제9권3호
• /
• pp.177-182
• /
• 2022

A deep neural network (DNN) includes variables whose values keep on changing with the training process until it reaches the final point of convergence. These variables are the co-efficient of a polynomial expression to relate to the feature extraction process. In general, DNNs work in multiple 'dimensions' depending upon the number of channels and batches accounted for training. However, after the execution of feature extraction and before entering the SoftMax or other classifier, there is a conversion of features from multiple N-dimensions to a single vector form, where 'N' represents the number of activation channels. This usually happens in a Fully connected layer (FCL) or a dense layer. This reduced 2D feature is the subject of study for our analysis. For this, we have used the FCL, so the trained weights of this FCL will be used for the weight-class correlation analysis. The popular DNN models selected for our study are ResNet-101, VGG-19, and GoogleNet. These models' weights are directly used for fine-tuning (with all trained weights initially transferred) and scratch trained (with no weights transferred). Then the comparison is done by plotting the graph of feature distribution and the final FCL weights.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2019년도 춘계학술발표대회
• /
• pp.472-474
• /
• 2019

Because of the difference in network structure and loss function, Verification and identification models have their respective advantages and limitations for person reidentification (re-ID). In this work, we propose a multi-task network simultaneously computes the identification loss and verification loss for person reidentification. Given a pair of images as network input, the multi-task network simultaneously outputs the identities of the two images and whether the images belong to the same identity. In experiments, we analyze the major factors affect the accuracy of person reidentification. To address the occlusion problem and improve the generalization ability of reID models, we use the Random Erasing Augmentation (REA) method to preprocess the images. The method can be easily applied to different pre-trained networks, such as ResNet and VGG. The experimental results on the Market1501 datasets show significant and consistent improvements over the state-of-the-art methods.

• KEPCO Journal on Electric Power and Energy
• /
• 제8권2호
• /
• pp.137-141
• /
• 2022

To inspect the power facility faults using artificial intelligence, it need that improve the accuracy of the diagnostic model are required. Data augmentation skill using generative adversarial network (GAN) is one of the best ways to improve deep learning performance. GAN model can create realistic-looking fake images using two competitive learning networks such as discriminator and generator. In this study, we intend to verify the effectiveness of virtual data generation technology by including the fake image of power facility generated through GAN in the deep learning training set. The GAN-based fake image was created for damage of LP insulator, and ResNet based normal and defect classification model was developed to verify the effect. Through this, we analyzed the model accuracy according to the ratio of normal and defective training data.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권12호
• /
• pp.3991-4007
• /
• 2022

In the public healthcare, a computational system that can automatically and efficiently detect and classify falls from a video sequence has significant potential. With the advancement of deep learning, which can extract temporal and spatial information, has become more widespread. However, traditional 3D CNNs that usually adopt shallow networks cannot obtain higher recognition accuracy than deeper networks. Additionally, some experiences of neural network show that the problem of gradient explosions occurs with increasing the network layers. As a result, an enhanced three-dimensional ResNet-based method for fall detection (3D-ERes-FD) is proposed to directly extract spatio-temporal features to address these issues. In our method, a 50-layer 3D residual network is used to deepen the network for improving fall recognition accuracy. Furthermore, enhanced residual units with four convolutional layers are developed to efficiently reduce the number of parameters and increase the depth of the network. According to the experimental results, the proposed method outperformed several state-of-the-art methods.

• Imaging Science in Dentistry
• /
• 제52권3호
• /
• pp.239-244
• /
• 2022

Purpose: Despite the proliferation of numerous morphometric and anthropometric methods for sex identification based on linear, angular, and regional measurements of various parts of the body, these methods are subject to error due to the observer's knowledge and expertise. This study aimed to explore the possibility of automated sex determination using convolutional neural networks(CNNs) based on lateral cephalometric radiographs. Materials and Methods: Lateral cephalometric radiographs of 1,476 Iranian subjects (794 women and 682 men) from 18 to 49 years of age were included. Lateral cephalometric radiographs were considered as a network input and output layer including 2 classes(male and female). Eighty percent of the data was used as a training set and the rest as a test set. Hyperparameter tuning of each network was done after preprocessing and data augmentation steps. The predictive performance of different architectures (DenseNet, ResNet, and VGG) was evaluated based on their accuracy in test sets. Results: The CNN based on the DenseNet121 architecture, with an overall accuracy of 90%, had the best predictive power in sex determination. The prediction accuracy of this model was almost equal for men and women. Furthermore, with all architectures, the use of transfer learning improved predictive performance. Conclusion: The results confirmed that a CNN could predict a person's sex with high accuracy. This prediction was independent of human bias because feature extraction was done automatically. However, for more accurate sex determination on a wider scale, further studies with larger sample sizes are desirable.

• Wind and Structures
• /
• 제36권6호
• /
• pp.405-421
• /
• 2023

Strong wind is the main factors of wind-damage of high-rise buildings, which often creates largely economical losses and casualties. Wind pressure plays a critical role in wind effects on buildings. To obtain the high-resolution wind pressure field, it often requires massive pressure taps. In this study, two traditional methods, including bilinear and bicubic interpolation, and two deep learning techniques including Residual Networks (ResNet) and Generative Adversarial Networks (GANs), are employed to reconstruct wind pressure filed from limited pressure taps on the surface of an ideal building from TPU database. It was found that the GANs model exhibits the best performance in reconstructing the wind pressure field. Meanwhile, it was confirmed that k-means clustering based retained pressure taps as model input can significantly improve the reconstruction ability of GANs model. Finally, the generalization ability of k-means clustering based GANs model in reconstructing wind pressure field is verified by an actual engineering structure. Importantly, the k-means clustering based GANs model can achieve satisfactory reconstruction in wind pressure field under the inputs processing by k-means clustering, even the 20% of pressure taps. Therefore, it is expected to save a huge number of pressure taps under the field reconstruction and achieve timely and accurately reconstruction of wind pressure field under k-means clustering based GANs model.