• 융합정보논문지
• /
• 제11권10호
• /
• pp.45-52
• /
• 2021

본 논문에서는 U-Net 기반의 semantic segmentation 방법에서 정확도를 개선하기 위한 Atrous Residual U-Net (AR-UNet)을 제안하였다. U-Net은 의료 영상 분석, 자율주행 자동차, 원격 감지 영상 등의 분야에서 주로 사용된다. 기존 U-Net은 인코더 부분에서 컨볼루션 계층 수가 적어 추출되는 특징이 부족하다. 추출된 특징은 객체의 범주를 분류하는 데 필수적이며, 부족할 경우 분할 정확도를 저하시키는 문제를 초래한다. 따라서 이 문제를 개선하기 위해 인코더에 residual learning과 ASPP를 활용한 AR-UNet을 제안하였다. Residual learning은 특징 추출 능력을 개선하고, 연속적인 컨볼루션으로 발생하는 특징 손실과 기울기 소실 문제 방지에 효과적이다. 또한 ASPP는 특징맵의 해상도를 줄이지 않고 추가적인 특징 추출이 가능하다. 실험은 Cityscapes 데이터셋으로 AR-UNet의 효과를 검증하였다. 실험 결과는 AR-UNet이 기존 U-Net과 비교하여 향상된 분할 결과를 보였다. 이를 통해 AR-UNet은 정확도가 중요한 여러 응용 분야의 발전에 기여할 수 있다.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송∙미디어공학회 2018년도 하계학술대회
• /
• pp.63-66
• /
• 2018

Recently, a convolutional neural network (CNN) models at single image super-resolution have been very successful. Residual learning improves training stability and network performance in CNN. In this paper, we compare four convolutional neural network models for super-resolution (SR) to learn nonlinear mapping from low-resolution (LR) input image to high-resolution (HR) target image. Four models include general CNN model, global residual learning CNN model, local residual learning CNN model, and the CNN model with global and local residual learning. Experiment results show that the results are greatly affected by how skip connections are connected at the basic CNN network, and network trained with only global residual learning generates highest performance among four models at objective and subjective evaluations.

• 디지털융복합연구
• /
• 제19권6호
• /
• pp.251-258
• /
• 2021

본 논문에서는 U-Net 기반의 semantic segmentation 방법에서 정확도를 향상시키기 위해 residual learning을 활용한 인코더-디코더 구조의 모델을 제안하였다. U-Net은 딥러닝 기반의 semantic segmentation 방법이며 자율주행 자동차, 의료 영상 분석과 같은 응용 분야에서 주로 사용된다. 기존 U-Net은 인코더의 얕은 구조로 인해 특징 압축 과정에서 손실이 발생한다. 특징 손실은 객체의 클래스 분류에 필요한 context 정보 부족을 초래하고 segmentation 정확도를 감소시키는 문제가 있다. 이를 개선하기 위해 제안하는 방법은 기존 U-Net에 특징 손실과 기울기 소실 문제를 방지하는데 효과적인 residual learning을 활용한 인코더를 통해 context 정보를 효율적으로 추출하였다. 또한, 인코더에서 down-sampling 연산을 줄여 특징맵에 포함된 공간 정보의 손실을 개선하였다. 제안하는 방법은 Cityscapes 데이터셋 실험에서 기존 U-Net 방법에 비해 segmentation 결과가 약 12% 향상되었다.

• Earthquakes and Structures
• /
• 제25권1호
• /
• pp.15-26
• /
• 2023

It is of great importance to assess the residual displacement demand in the performance-based seismic design. In this paper, a hybrid deep learning model for predicting the residual displacement spectra under near-fault (NF) ground motions is proposed by combining the long short-term memory network (LSTM) and back-propagation (BP) network. The model is featured by its capacity of predicting the residual displacement spectrum under a given NF ground motion while considering the effects of structural parameters. To construct this model, 315 natural and artificial NF ground motions were employed to compute the residual displacement spectra through elastoplastic time history analysis considering different structural parameters. Based on the resulted dataset with a total of 9,450 samples, the proposed model was finally trained and tested. The results show that the proposed model has a satisfactory accuracy as well as a high efficiency in predicting residual displacement spectra under given NF ground motions while considering the impacts of structural parameters.

• 한국멀티미디어학회논문지
• /
• 제24권1호
• /
• pp.67-74
• /
• 2021

As IoT(Internet Of Things) devices like a smart sensor have constrained power sources, a power strategy is critical in WSN(Wireless Sensor Networks). Therefore, it is necessary to figure out the residual power of each sensor node for managing power strategies in WSN, which, however, requires additional data transmission, leading to more power consumption. In this paper, a residual power estimation method was proposed, which uses ignorantly small amount of power consumption in the resource-constrained wireless networks including WSN. A residual power prediction is possible with the least data transmission by using Machine Learning method with some training data in this proposal. The performance of the proposed scheme was evaluated by machine learning method, simulation, and analysis.

• Journal of Information Processing Systems
• /
• 제17권2호
• /
• pp.385-398
• /
• 2021

The complexity of deep learning models affects the real-time performance of gesture recognition, thereby limiting the application of gesture recognition algorithms in actual scenarios. Hence, a residual learning neural network based on a deep convolutional neural network is proposed. First, small convolution kernels are used to extract the local details of gesture images. Subsequently, a shallow residual structure is built to share weights, thereby avoiding gradient disappearance or gradient explosion as the network layer deepens; consequently, the difficulty of model optimisation is simplified. Additional convolutional neural networks are used to accelerate the refinement of deep abstract features based on the spatial importance of the gesture feature distribution. Finally, a fully connected cascade softmax classifier is used to complete the gesture recognition. Compared with the dense connection multiplexing feature information network, the proposed algorithm is optimised in feature multiplexing to avoid performance fluctuations caused by feature redundancy. Experimental results from the ISOGD gesture dataset and Gesture dataset prove that the proposed algorithm affords a fast convergence speed and high accuracy.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송∙미디어공학회 2019년도 하계학술대회
• /
• pp.98-101
• /
• 2019

At present, deep convolutional neural networks have made a very important contribution in single-image super-resolution. Through the learning of the neural networks, the features of input images are transformed and combined to establish a nonlinear mapping of low-resolution images to high-resolution images. Some previous methods are difficult to train and take up a lot of memory. In this paper, we proposed a simple and compact deep recursive residual network learning the features for single image super resolution. Global residual learning and local residual learning are used to reduce the problems of training deep neural networks. And the recursive structure controls the number of parameters to save memory. Experimental results show that the proposed method improved image qualities that occur in previous methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권9호
• /
• pp.2882-2903
• /
• 2022

Image super-resolution (SR) processing is of great value in the fields of digital image processing, intelligent security, film and television production and so on. This paper proposed a densely connected deep learning network based on cascade architecture, which can be used to solve the problem of super-resolution in the field of image quality enhancement. We proposed a more efficient residual scaling dense block (RSDB) and the multi-channel cascade architecture to realize more efficient feature reuse. Also we proposed a hybrid loss function based on L₁ error and L_∞ error to achieve better L_∞ error performance. The experimental results show that the overall performance of the network is effectively improved on cascade architecture and residual scaling. Compared with the residual dense net (RDN), the PSNR / SSIM of the new method is improved by 2.24% / 1.44% respectively, and the L_∞ performance is improved by 3.64%. It shows that the cascade connection and residual scaling method can effectively realize feature reuse, improving the residual convergence speed and learning efficiency of our network. The L_∞ performance is improved by 11.09% with only a minimal loses of 1.14% / 0.60% on PSNR / SSIM performance after adopting the new loss function. That is to say, the L_∞ performance can be improved greatly on the new loss function with a minor loss of PSNR / SSIM performance, which is of great value in L_∞ error sensitive tasks.

• International journal of advanced smart convergence
• /
• 제12권4호
• /
• pp.142-146
• /
• 2023

We present a method for generating 3D structures and rendering objects by combining VAE (Variational Autoencoder) and GAN (Generative Adversarial Network). This approach focuses on generating and rendering 3D models with improved quality using residual learning as the learning method for the encoder. We deep stack the encoder layers to accurately reflect the features of the image and apply residual blocks to solve the problems of deep layers to improve the encoder performance. This solves the problems of gradient vanishing and exploding, which are problems when constructing a deep neural network, and creates a 3D model of improved quality. To accurately extract image features, we construct deep layers of the encoder model and apply the residual function to learning to model with more detailed information. The generated model has more detailed voxels for more accurate representation, is rendered by adding materials and lighting, and is finally converted into a mesh model. 3D models have excellent visual quality and accuracy, making them useful in various fields such as virtual reality, game development, and metaverse.

• Journal of Information Processing Systems
• /
• 제17권6호
• /
• pp.1115-1126
• /
• 2021

Plant disease is one of the most irritating problems for agriculture growers. Thus, timely detection of plant diseases is of high importance to practical value, and corresponding measures can be taken at the early stage of plant diseases. Therefore, numerous researchers have made unremitting efforts in plant disease identification. However, this problem was not solved effectively until the development of artificial intelligence and big data technologies, especially the wide application of deep learning models in different fields. Since the symptoms of plant diseases mainly appear visually on leaves, computer vision and machine learning technologies are effective and rapid methods for identifying various kinds of plant diseases. As one of the fruits with the highest nutritional value, apple production directly affects the quality of life, and it is important to prevent disease intrusion in advance for yield and taste. In this study, an improved deep residual network is proposed for apple leaf disease identification in a novel way, a global residual connection is added to the original residual network, and the local residual connection architecture is optimized. Including that 1,977 apple leaf disease images with three categories that are collected in this study, experimental results show that the proposed method has achieved 98.74% top-1 accuracy on the test set, outperforming the existing state-of-the-art models in apple leaf disease identification tasks, and proving the effectiveness of the proposed method.