• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권5호
• /
• pp.1778-1797
• /
• 2021

Document image binarization is an important pre-processing step in document analysis and archiving. The state-of-the-art models for document image binarization are variants of encoder-decoder architectures, such as FCN (fully convolutional network) and U-Net. Despite their success, they still suffer from three limitations: (1) reduced feature map resolution due to consecutive strided pooling or convolutions, (2) multiple scales of target objects, and (3) reduced localization accuracy due to the built-in invariance of deep convolutional neural networks (DCNNs). To overcome these three challenges, we propose an improved semantic segmentation model, referred to as DP-LinkNet, which adopts the D-LinkNet architecture as its backbone, with the proposed hybrid dilated convolution (HDC) and spatial pyramid pooling (SPP) modules between the encoder and the decoder. Extensive experiments are conducted on recent document image binarization competition (DIBCO) and handwritten document image binarization competition (H-DIBCO) benchmark datasets. Results show that our proposed DP-LinkNet outperforms other state-of-the-art techniques by a large margin. Our implementation and the pre-trained models are available at https://github.com/beargolden/DP-LinkNet.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권1호
• /
• pp.180-194
• /
• 2021

For small image registration, feature-based approaches are likely to fail as feature detectors cannot detect enough feature points from low-resolution images. The classic FFT approach's prediction accuracy is high, but the registration time can be relatively long, about several seconds to register one image pair. To achieve real-time and high-precision rigid registration for small images, we apply deep neural networks for supervised rigid transformation prediction, which directly predicts the transformation parameters. We train deep registration models with rigidly transformed CIFAR-10 images and STL-10 images, and evaluate the generalization ability of deep registration models with transformed CIFAR-10 images, STL-10 images, and randomly generated images. Experimental results show that the deep registration models we propose can achieve comparable accuracy to the classic FFT approach for small CIFAR-10 images (32×32) and our LSTM registration model takes less than 1ms to register one pair of images. For moderate size STL-10 images (96×96), FFT significantly outperforms deep registration models in terms of accuracy but is also considerably slower. Our results suggest that deep registration models have competitive advantages over conventional approaches, at least for small images.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제11권4호
• /
• pp.119-126
• /
• 2022

본 연구는 폐쇄망에서 효율적이고 안전하게 패키지 관리 시스템을 사용하기 위해서 고려해야 할 할 중요 요소들과 그 방법론 들을 제시하는 것을 목적으로 한다. 관련 선행 연구의 분석을 통해 기존 패키지 관리에서 보안성을 위해 고려해야 할 사항들을 살펴보고, 이를 바탕으로 폐쇄망이라는 특수한 상황에서 고려해야 할 세부 방법들을 제안한다. 구체적으로, 새로운 패키지 관리 도구의 개발, 물리적 저장매체 활용, 로컬 백업 저장소 활용, 패키지 업데이트 및 다운그레이드 일괄 처리의 방법을 제안한다.

• 스마트미디어저널
• /
• 제11권4호
• /
• pp.38-45
• /
• 2022

Object detection models have become the current tool of choice for plant disease detection in precision agriculture. Most existing research improves the performance by ameliorating networks and optimizing the loss function. However, the data-centric part of a whole project also needs more investigation. In this paper, we proposed a systematic strategy with three different annotation methods for plant disease detection: local, semi-global, and global label. Experimental results on our paprika disease dataset show that a single class annotation with semi-global boxes may improve accuracy. In addition, we also studied the noise factor during the labeling process. An ablation study shows that annotation noise within 10% is acceptable for keeping good performance. Overall, this data-centric numerical analysis helps us to understand the significance of annotation methods, which provides practitioners a way to obtain higher performance and reduce annotation costs on plant disease detection tasks. Our work encourages researchers to pay more attention to label quality and the essential issues of labeling methods.

• 한국멀티미디어학회논문지
• /
• 제25권6호
• /
• pp.878-885
• /
• 2022

Research in the field of computer vision based on deep learning is being actively conducted. However, deep learning-based models have vulnerabilities in adversarial attacks that increase the model's misclassification rate by applying adversarial perturbation. In particular, in the case of FGSM, it is recognized as one of the effective attack methods because it is simple, fast and has a considerable attack success rate. Meanwhile, as one of the efforts to visualize deep learning models, Grad-CAM enables visual explanation of convolutional neural networks. In this paper, I propose a method to generate adversarial examples with high attack success rate by applying Grad-CAM to FGSM. The method chooses fixels, which are closely related to labels, by using Grad-CAM and add perturbations to the fixels intensively. The proposed method has a higher success rate than the FGSM model in the same perturbation for both targeted and untargeted examples. In addition, unlike FGSM, it has the advantage that the distribution of noise is not uniform, and when the success rate is increased by repeatedly applying noise, the attack is successful with fewer iterations.

• Wind and Structures
• /
• 제34권6호
• /
• pp.511-523
• /
• 2022

Detecting the icing on wind turbine blades built-in cold regions with conventional methods is always a very laborious, expensive and very difficult task. Regarding this issue, the use of smart systems has recently come to the agenda. It is quite possible to eliminate this issue by using the deep learning method, which is one of these methods. In this study, an application has been implemented that can detect icing on wind turbine blades images with visualization techniques based on deep learning using images. Pre-trained models of Resnet-50, VGG-16, VGG-19 and Inception-V3, which are well-known deep learning approaches, are used to classify objects automatically. Grad-CAM, Grad-CAM++, and Score-CAM visualization techniques were considered depending on the deep learning methods used to predict the location of icing regions on the wind turbine blades accurately. It was clearly shown that the best visualization technique for localization is Score-CAM. Finally, visualization performance analyses in various cases which are close-up and remote photos of a wind turbine, density of icing and light were carried out using Score-CAM for Resnet-50. As a result, it is understood that these methods can detect icing occurring on the wind turbine with acceptable high accuracy.

• Journal of information and communication convergence engineering
• /
• 제20권4호
• /
• pp.250-258
• /
• 2022

The emerging scope of the Internet-of-Things (IoT) has piqued the interest of industry and academia in recent times. Therefore, security becomes the main issue to prevent the possibility of cyberattacks. Jamming attacks are threads that can affect performance and cause significant problems for IoT device. This study explores a smart jamming attack (coalition attack) in which the attackers were previously a part of the legitimate network and are now back to attack it based on the gained knowledge. These attackers regroup into a coalition and begin exchanging information about the legitimate network to launch attacks based on the gained knowledge. Our system enables jammer nodes to select the optimal transmission rates for attacks based on the attack probability table, which contains the most probable link transmission rate between nodes in the legitimate network. The table is updated constantly throughout the life cycle of the coalition. The simulation results show that a coalition of jammers can cause highly successful attacks.

• 대한임베디드공학회논문지
• /
• 제17권6호
• /
• pp.367-374
• /
• 2022

The acceleration of neural networks has become an important topic in the field of computer vision. An accelerator is absolutely necessary for accelerating the lightweight model. Most accelerator-supported operators focused on direct convolution operations. If the accelerator does not provide GEMM operation, it is mostly replaced by CPU operation. In this paper, we proposed an optimization technique for 2-stage tiling-based GEMM routines on VTA. We improved performance of the matrix multiplication routine by maximizing the reusability of the input matrix and optimizing the operation pipelining. In addition, we applied the proposed technique to the DarkNet framework to check the performance improvement of the matrix multiplication routine. The proposed GEMM method showed a performance improvement of more than 2.4 times compared to the non-optimized GEMM method. The inference performance of our DarkNet framework has also improved by at least 2.3 times.

• 한국인터넷방송통신학회논문지
• /
• 제22권3호
• /
• pp.69-74
• /
• 2022

뉴로모픽 아키텍처에서 동작하는 SNN (Spiking Neural Network) 은 인간의 신경망을 모방하여 만들어졌다. 뉴로모픽 아키텍처 기반의 뉴로모픽 컴퓨팅은 GPU를 이용한 딥러닝 기법보다 상대적으로 낮은 전력을 요구한다. 이와 같은 이유로 뉴로모픽 아키텍처를 이용하여 다양한 인공지능 모델을 지원하고자 하는 연구가 활발히 일어나고 있다. 본 논문에서는 음성 데이터 전처리 기법에 따른 뉴로모픽 아키텍처 기반의 음성 인식 모델의 성능 분석을 진행하였다. 실험 결과 푸리에 변환 기반 음성 데이터 전처리시 최대 84% 정도의 인식 정확도 성능을 보임을 확인하였다. 따라서 뉴로모픽 아키텍처 기반의 음성 인식 서비스가 효과적으로 활용될 수 있음을 확인하였다.

• 한국전자통신학회논문지
• /
• 제16권6호
• /
• pp.1113-1120
• /
• 2021

콘텐츠 중심 네트워크(CCN)은 기존 위치(IP) 기반의 네트워크 방식을 콘텐츠 이름(Content Name) 중심의 네트워크 구조로 변모시킴으로써 네트워크 전반의 효율성을 높이고자 하는 시도이다. CCN에서는 네트워크 효율을 높이기 위해 라우터 저장공간을 활용한 캐싱을 수행하는데, 캐시 교체정책은 CCN의 전반적인 성능을 좌우하는 중요한 요소이다. 따라서 CCN 분야에서는 캐시 교체정책과 관련된 많은 선행 연구가 있었다. 본 논문에서는 CCN 기본 캐시 교체정책인 LFU를 개선한 실시간 생성자 인기도 기반의 캐시 교체정책을 제안하였다. 또한, 실험을 통해 제안한 캐시 교체정책이 대조군보다 우수함을 입증하였다.