• Structural Monitoring and Maintenance
• /
• 제8권4호
• /
• pp.379-402
• /
• 2021

Traditional approaches for structural health monitoring (SHM) seldom take ambient uncertainty (temperature, humidity, ambient vibration) into consideration, while their impacts on structural responses are substantial, leading to a possibility of raising false alarms. A few predictors model-based approaches deal with these uncertainties through complex numerical models running online, rendering the SHM approach to be compute-intensive, slow, and sometimes not practical. Also, with model-based approaches, the imperative need for a precise understanding of the structure often poses a problem for not so well understood complex systems. The present study employs a data-based approach coupled with Empirical mode decomposition (EMD) to correlate recorded response time histories under varying temperature conditions to corresponding damage scenarios. EMD decomposes the response signal into a finite set of intrinsic mode functions (IMFs). A two-dimensional Convolutional Neural Network (2DCNN) is further trained to associate these IMFs to the respective damage cases. The use of IMFs in place of raw signals helps to reduce the impact of sensor noise while preserving the essential spatio-temporal information less-sensitive to thermal effects and thereby stands as a better damage-sensitive feature than the raw signal itself. The proposed algorithm is numerically tested on a single span bridge under varying temperature conditions for different damage severities. The dynamic strain is recorded as the response since they are frame-invariant and cheaper to install. The proposed algorithm has been observed to be damage sensitive as well as sufficiently robust against measurement noise.

• 한국멀티미디어학회논문지
• /
• 제24권12호
• /
• pp.1598-1605
• /
• 2021

Recently, convolutional neural networks (CNNs) have been widely used with excellent performance in various computer vision fields, including super-resolution (SR). However, CNN is computationally intensive and requires a lot of memory, making it difficult to apply to limited hardware resources such as mobile or Internet of Things devices. To solve these limitations, network lightening studies have been actively conducted to reduce the depth or size of pre-trained deep CNN models while maintaining their performance as much as possible. This paper aims to lighten the SR CNN model, SRGAN, using the knowledge distillation among network lightening technologies; thus, it proposes four techniques with different methods of transferring the knowledge of the teacher network to the student network and presents experiments to compare and analyze the performance of each technique. In our experimental results, it was confirmed through quantitative and qualitative evaluation indicators that student networks with knowledge transfer performed better than those without knowledge transfer, and among the four knowledge transfer techniques, the technique of conducting adversarial learning after transferring knowledge from the teacher generator to the student generator showed the best performance.

• 한국멀티미디어학회논문지
• /
• 제24권11호
• /
• pp.1552-1559
• /
• 2021

Convolution neural networks (CNNs) show notable performance in image processing and are used as representative core models. CNNs extract and learn features from large amounts of train dataset. In general, it has a structure in which a convolution layer and a fully connected layer are stacked. The core of CNN is the convolution layer. The size of the kernel used for feature extraction and the number that affect the depth of the feature map determine the amount of weight parameters of the CNN that can be learned. These parameters are the main causes of increasing the computational complexity and memory usage of the entire neural network. The most computationally expensive components in CNNs are fully connected and spatial convolution computations. In this paper, we propose a Fourier Convolution Neural Network that performs the operation of the convolution layer in the Fourier domain. We work on modifying and improving the amount of computation by applying the fast fourier transform method. Using the MNIST dataset, the performance was similar to that of the general CNN in terms of accuracy. In terms of operation speed, 7.2% faster operation speed was achieved. An average of 19% faster speed was achieved in experiments using 1024x1024 images and various sizes of kernels.

• International Journal of Computer Science & Network Security
• /
• 제21권11호
• /
• pp.73-80
• /
• 2021

Nowadays microblogs have become the most popular platforms to obtain and spread information. Twitter is one of the most used platforms to share everyday life event. However, rumors and misinformation on Arabic social media platforms has become pervasive which can create inestimable harm to society. Therefore, it is imperative to tackle and study this issue to distinguish the verified information from the unverified ones. There is an increasing interest in rumor detection on microblogs recently, however, it is mostly applied on English language while the work on Arabic language is still ongoing research topic and need more efforts. In this paper, we propose a combined Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) to detect rumors on Twitter dataset. Various experiments were conducted to choose the best hyper-parameters tuning to achieve the best results. Moreover, different neural network models are used to evaluate performance and compare results. Experiments show that the CNN-LSTM model achieved the best accuracy 0.95 and an F1-score of 0.94 which outperform the state-of-the-art methods.

• Smart Structures and Systems
• /
• 제29권1호
• /
• pp.153-166
• /
• 2022

Identifying fine cracks in steel bridge facilities is a challenging task of structural health monitoring (SHM). This study proposed an end-to-end crack image segmentation framework based on a one-step Convolutional Neural Network (CNN) for pixel-level object recognition with high accuracy. To particularly address the challenges arising from small object detection in complex background, efforts were made in loss function selection aiming at sample imbalance and module modification in order to improve the generalization ability on complicated images. Specifically, loss functions were compared among alternatives including the Binary Cross Entropy (BCE), Focal, Tversky and Dice loss, with the last three specialized for biased sample distribution. Structural modifications with dilated convolution, Spatial Pyramid Pooling (SPP) and Feature Pyramid Network (FPN) were also performed to form a new backbone termed CrackDet. Models of various loss functions and feature extraction modules were trained on crack images and tested on full-scale images collected on steel box girders. The CNN model incorporated the classic U-Net as its backbone, and Dice loss as its loss function achieved the highest mean Intersection-over-Union (mIoU) of 0.7571 on full-scale pictures. In contrast, the best performance on cropped crack images was achieved by integrating CrackDet with Dice loss at a mIoU of 0.7670.

• International Journal of Computer Science & Network Security
• /
• 제23권9호
• /
• pp.47-54
• /
• 2023

For Bengali music emotion classification, deep learning models, particularly CNN and RNN are frequently used. But previous researches had the flaws of low accuracy and overfitting problem. In this research, attention-based Conv1D and BiGRU model is designed for music emotion classification and comparative experimentation shows that the proposed model is classifying emotions more accurate. We have proposed a Conv1D and Bi-GRU with the attention-based model for emotion classification of our Bengali music dataset. The model integrates attention-based. Wav preprocessing makes use of MFCCs. To reduce the dimensionality of the feature space, contextual features were extracted from two Conv1D layers. In order to solve the overfitting problems, dropouts are utilized. Two bidirectional GRUs networks are used to update previous and future emotion representation of the output from the Conv1D layers. Two BiGRU layers are conntected to an attention mechanism to give various MFCC feature vectors more attention. Moreover, the attention mechanism has increased the accuracy of the proposed classification model. The vector is finally classified into four emotion classes: Angry, Happy, Relax, Sad; using a dense, fully connected layer with softmax activation. The proposed Conv1D+BiGRU+Attention model is efficient at classifying emotions in the Bengali music dataset than baseline methods. For our Bengali music dataset, the performance of our proposed model is 95%.

• Journal of Platform Technology
• /
• 제11권1호
• /
• pp.72-84
• /
• 2023

With the rise of the Internet of Things, the security of such lightweight computing environments has become a hot topic. Lightweight block ciphers that can provide efficient performance and security by having a relatively simpler structure and smaller key and block sizes are drawing attention. Due to these characteristics, they can become a target for new attack techniques. One of the new cryptanalytic attacks that have been attracting interest is Neural cryptanalysis, which is a cryptanalytic technique based on neural networks. It showed interesting results with better results than the conventional cryptanalysis method without a great amount of time and cryptographic knowledge. The first work that showed good results was carried out by Aron Gohr in CRYPTO'19, the attack was conducted on the lightweight block cipher SPECK-/32/64 and showed better results than conventional differential cryptanalysis. In this paper, we first apply the Differential Neural Distinguisher proposed by Aron Gohr to the block ciphers HIGHT and GOST to test the applicability of the attack to ciphers with different structures. The performance of the Differential Neural Distinguisher is then analyzed by replacing the neural network attack model with five different models (Multi-Layer Perceptron, AlexNet, ResNext, SE-ResNet, SE-ResNext). We then propose a Related-key Neural Distinguisher and apply it to the SPECK-/32/64, HIGHT, and GOST block ciphers. The proposed Related-key Neural Distinguisher was constructed using the relationship between keys, and this made it possible to distinguish more rounds than the differential distinguisher.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2022년도 학술발표회
• /
• pp.131-131
• /
• 2022

기후변화의 영향으로 극치강우의 변동성이 커지고 있으며 계획빈도를 초과하는 폭우로 피해가 증가하고 있다. 기존의 물리기반의 홍수예측모델은 개념적 및 구조적 제약과 함께 다양한 유역조건 및 수문기상 조건에 기인한 강우-유출 관계의 불확실성을 고려하는 데 한계가 있다. 특히 한정된 홍수 사상을 통해 구축된 관측 자료로 인해 새로운 홍수 사상 예측 능력이 저조할 수밖에 없다. 따라서 기존 물리모형 기반의 홍수예측과 함께, 딥러닝(deep learning) 모형을 고려한 홍수예측 모델 개발과 개선이 필요하다. 본 연구에서는 다양한 분야에서 활용되는 인공지능(artificial intelligence, AI) 기술을 종합적으로 검토하고, 홍수 예측 측면에서의 활용 가능성 및 신뢰성을 고려하여 AI 기법을 채택하였다. 한강수계에 존재하는 댐 중 일부를 선정하여 대상 댐의 수문·기상학적 자료를 전처리한 후, 인공지능 기반의 홍수예측모형을 구축 및 최적화하였다. 다양한 예측인자와 모델 구성으로 홍수예측력에 대한 평가를 다각적으로 수행함으로써 홍수예측모델의 신뢰성을 제고하였다. 전반적으로 우수한 결과를 도출하였고, 유역면적이 작을수록 결과가 좋았다. 이는 넓은 유역일수록 복잡한 강우-유출 과정이 내재되어 있기 때문으로 판단되며, 넓은 유역에는 본 연구에서 활용한 자료에 추가적인 자료를 도입하여 모형 개선이 이루어져야 할 것으로 판단하였다. 수문 예측 연구에 통계모형이나 기계학습모형의 적용은 많이 있었지만, 딥러닝 기법 활용은 새로운 시도라는 점에서 의미가 있다.

• 대한임베디드공학회논문지
• /
• 제18권2호
• /
• pp.41-50
• /
• 2023

Gait analysis is an important tool in the clinical management of cerebral palsy, allowing for the assessment of condition severity, identification of potential gait abnormalities, planning and evaluation of interventions, and providing a baseline for future comparisons. However, traditional methods of gait analysis are costly and time-consuming, leading to a need for a more convenient and continuous method. This paper proposes a method for analyzing the posture of cerebral palsy patients using only smartphone videos and deep learning models, including a ResNet-based image tilt correction, AlphaPose for human pose estimation, and SmoothNet for temporal smoothing. The indicators employed in medical practice, such as the imbalance angles of shoulder and pelvis and the joint angles of spine-thighs, knees and ankles, were precisely examined. The proposed system surpassed pose estimation alone, reducing the mean absolute error for imbalance angles in frontal videos from 4.196° to 2.971° and for joint angles in sagittal videos from 5.889° to 5.442°.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제11권2호
• /
• pp.93-100
• /
• 2022

기존의 데이터 검색 방법으로는 키워드 중심의 검색 방법이 주로 사용되나, 이는 전문적인 용어가 많이 쓰이는 법률 분야의 검색 방법으로는 적합하지 않다. 이에 대해 본 논문에서는 법률 분야의 효과적인 데이터 검색 방안을 제안한다. 법률 도메인의 자연어처리 분야에서 문장 간의 유사성을 판단하는 데 최적화된 임베딩 방법에 관하여 서술한다. 법률문장을 TF-IDF를 이용하여 키워드 기반으로 임베딩하거나 Universal Sentence Encoder를 이용하여 의미 기반으로 임베딩을 한 후, BERT모델을 결합하여 법률 분야에서 문장 간 유사성을 검사하여 데이터를 검색하는 최적의 방안을 제안한다.