• Journal of Korea Multimedia Society
• /
• v.25 no.11
• /
• pp.1582-1590
• /
• 2022

Due to the situation of the widespread of the coronavirus, which causes the problem of lack of face image data occluded by masks at recent time, in order to solve the related problems, this paper proposes a method to generate face images with masks using a combination of generative adversarial networks and spatial transformation networks based on CNN model. The system we proposed in this paper is based on the GAN, combined with multi-scale convolution kernels to extract features at different details of the human face images, and used Wasserstein divergence as the measure of the distance between real samples and synthetic samples in order to optimize Generator performance. Experiments show that the proposed method can effectively put masks on face images with high efficiency and fast reaction time and the synthesized human face images are pretty natural and real.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.1
• /
• pp.1-15
• /
• 2022

The MobileNetV3 is specially designed for mobile devices with limited memory and computing power. To reduce the network parameters and improve the network inference speed, a new lightweight network is proposed based on MobileNetV3. Firstly, to reduce the computation of residual blocks, a partial residual structure is designed by dividing the input feature maps into two parts. The designed partial residual structure is used to replace the residual block in MobileNetV3. Secondly, a dual-path feature extraction structure is designed to further reduce the computation of MobileNetV3. Different convolution kernel sizes are used in the two paths to extract feature maps with different sizes. Besides, a transition layer is also designed for fusing features to reduce the influence of the new structure on accuracy. The CIFAR-100 dataset and Image Net dataset are used to test the performance of the proposed partial residual structure. The ResNet based on the proposed partial residual structure has smaller parameters and FLOPs than the original ResNet. The performance of improved MobileNetV3 is tested on CIFAR-10, CIFAR-100 and ImageNet image classification task dataset. Comparing MobileNetV3, GhostNet and MobileNetV2, the improved MobileNetV3 has smaller parameters and FLOPs. Besides, the improved MobileNetV3 is also tested on CPU and Raspberry Pi. It is faster than other networks

• Journal of Broadcast Engineering
• /
• v.25 no.2
• /
• pp.242-251
• /
• 2020

Recently, many deep convolutional neural networks for image super-resolution have been studied. Existing deep learning-based super-resolution algorithms are architecture that up-samples the resolution at the end of the network. The post-upsampling architecture has an inefficient structure at large scaling factor result of predicting a lot of information for mapping from low-resolution to high-resolution at once. In this paper, we propose a single image super-resolution using Channel Attention Residual Dense Block based on an iterative up-down sampling architecture. The proposed algorithm efficiently predicts the mapping relationship between low-resolution and high-resolution, and shows up to 0.14dB performance improvement and enhanced subjective image quality compared to the existing algorithm at large scaling factor result.

• International Journal of Computer Science & Network Security
• /
• v.22 no.10
• /
• pp.73-82
• /
• 2022

Early detection continues to be the mainstay of breast cancer control as well as the improvement of its treatment. Even so, the absence of cancer symptoms at the onset has early detection quite challenging. Therefore, various researchers continue to focus on cancer as a topic of health to try and make improvements from the perspectives of diagnosis, prevention, and treatment. This research's chief goal is development of a system with deep learning for classification of the breast cancer as non-malignant and malignant using mammogram images. The following two distinct approaches: the first one with the utilization of patches of the Region of Interest (ROI), and the second one with the utilization of the overall images is used. The proposed system is composed of the following two distinct stages: the pre-processing stage and the Convolution Neural Network (CNN) building stage. Of late, the use of meta-heuristic optimization algorithms has accomplished a lot of progress in resolving these problems. Teaching-Learning Based Optimization algorithm (TIBO) meta-heuristic was originally employed for resolving problems of continuous optimization. This work has offered the proposals of novel methods for training the Residual Network (ResNet) as well as the CNN based on the TLBO and the Genetic Algorithm (GA). The classification of breast cancer can be enhanced with direct application of the hybrid TLBO- GA. For this hybrid algorithm, the TLBO, i.e., a core component, will combine the following three distinct operators of the GA: coding, crossover, and mutation. In the TLBO, there is a representation of the optimization solutions as students. On the other hand, the hybrid TLBO-GA will have further division of the students as follows: the top students, the ordinary students, and the poor students. The experiments demonstrated that the proposed hybrid TLBO-GA is more effective than TLBO and GA.

• Geophysics and Geophysical Exploration
• /
• v.23 no.3
• /
• pp.117-130
• /
• 2020

Noises can distort acquired geophysical data, leading to their misinterpretation. Potential noises sources include anthropogenic activity, natural phenomena, and instrument noises. Conventional denoising methods such as wavelet transform and filtering techniques, are based on subjective human investigation, which is computationally inefficient and time-consuming. Recently, many researchers attempted to implement neural networks to efficiently remove noise from geophysical data. This study aims to review and analyze different types of neural networks, such as artificial neural networks, convolutional neural networks, autoencoders, residual networks, and wavelet neural networks, which are implemented to remove different types of noises including seismic, transient electromagnetic, ground-penetrating radar, and magnetotelluric surveys. The review analyzes and summarizes the key challenges in the removal of noise from geophysical data using neural network, while proposes and explains solutions to the challenges. The analysis support that the advancement in neural networks can be powerful denoising tools for geophysical data.

• Smart Media Journal
• /
• v.9 no.1
• /
• pp.23-29
• /
• 2020

This paper presents an approach for dynamic hand gesture recognition by using algorithm based on 3D Convolutional Neural Network (3D_CNN), which is later extended to 3D Residual Networks (3D_ResNet), and the neural network based key frame selection. Typically, 3D deep neural network is used to classify gestures from the input of image frames, randomly sampled from a video data. In this work, to improve the classification performance, we employ key frames which represent the overall video, as the input of the classification network. The key frames are extracted by SegNet instead of conventional clustering algorithms for video summarization (VSUMM) which require heavy computation. By using a deep neural network, key frame selection can be performed in a real-time system. Experiments are conducted using 3D convolutional kernels such as 3D_CNN, Inflated 3D_CNN (I3D) and 3D_ResNet for gesture classification. Our algorithm achieved up to 97.8% of classification accuracy on the Cambridge gesture dataset. The experimental results show that the proposed approach is efficient and outperforms existing methods.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.19 no.12
• /
• pp.21-27
• /
• 2020

Lithium-ion batteries are the heart of energy-storing devices and electric vehicles. Owing to their superior qualities, such as high capacity and energy efficiency, they have become quite popular, resulting in an increased demand for failure/damage prevention and useable life maximization. To prevent failure in Lithium-ion batteries, improve their reliability, and ensure productivity, prognosticative measures such as condition monitoring through sensors, condition assessment for failure detection, and remaining useful life prediction through data-driven prognostics and health management approaches have become important topics for research. In this study, the residual useful life of Lithium-ion batteries was predicted using two efficient artificial recurrent neural networks-ong short-term memory (LSTM) and gated recurrent unit (GRU). The proposed approaches were compared for prognostics accuracy and cost-efficiency. It was determined that LSTM showed slightly higher accuracy, whereas GRUs have a computational advantage.

• Journal of Communications and Networks
• /
• v.8 no.2
• /
• pp.212-219
• /
• 2006

In this paper, a new turbo codes-based hybrid automatic repeat request (TC-HARQ) scheme with segment selective repeat (SSR) is proposed. The main strategy is, upon retransmission, to repeat the data that are most important for the next round of decoding based on the distribution of residual errors after current decoding. The performance in terms of reliability and throughput is analyzed. To adapt to correlated fading channels where an inter-leaver is always employed before transmission, we further modify the SSR strategy so that data having experienced correlated deep fading are selected for retransmission. Finally, this proposed scheme is applied to the wideband code division multiple access (WCDMA) system under frequency selective fading channels. Simulation results demonstrate that in all single and multiple user cases, SSR-based TC-HARQ leads to significant throughput improvement with similar bit error rate (BER) performance as compared to type-I TC-HARQ.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.11a
• /
• pp.62-65
• /
• 2018

신경망은 깊어질수록 gradient vanishing/exploding과 같은 네트워크가 불안정해지는 문제가 발생 한다. 잔차 블록을 이용하여 이러한 문제를 해결 할 수 있다. 본 논문에서는 영상 인식 분야에서 훌륭한 성능을 보여준 잔차 블록 기반의 깊은 합성곱 신경망을 통한 단일 영상 초해상도 복원 기법을 제안 한다. 제안한 알고리듬은 EDSR에 사용된 잔차 블록을 다양한 크기의 합성곱 연산을 통해 영상의 특징들을 다르게 분석하도록 수정하고 VDSR과 비슷한 수준의 복잡도로 구성하여 향상된 성능을 얻었다. 실험 결과, VDSR에 비해 PSNR이 최대 0.1dB까지 증가했다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• fall
• /
• pp.80-81
• /
• 2021

딥 러닝 모델 사용에 있어서, 일반적인 사용자가 이용할 수 있는 하드웨어 리소스는 제한적이기 때문에 기존 모델을 경량화 할 수 있는 프루닝 방법을 통해 제한적인 리소스를 효과적으로 활용할 수 있도록 한다. 그 방법으로, 여러 딥 러닝 모델들 중 비교적 파라미터 수가 많은 것으로 알려진 GAN 아키텍처에 네트워크 프루닝을 적용함으로써 비교적 무거운 모델을 적은 파라미터를 통해 학습할 수 있는 방법을 제시한다. 또한, 본 논문을 통해 기존의 SRGAN 논문에서 가장 효과적인 결과로 제시했던 16 개의 residual block 의 개수를 실제로 줄여 봄으로써 기존 논문에서 제시했던 결과와의 차이에 대해 서술한다.