• The Journal of the Acoustical Society of Korea
• /
• v.38 no.6
• /
• pp.670-677
• /
• 2019

As the presence of background noise in acoustic signal degrades the performance of speech or acoustic event recognition, it is still challenging to extract noise-robust acoustic features from noisy signal. In this paper, we propose a combined structure of Wasserstein Generative Adversarial Network (WGAN) and MultiTask AutoEncoder (MTAE) as deep learning architecture that integrates the strength of MTAE and WGAN respectively such that it estimates not only noise but also speech features from noisy acoustic source. The proposed MTAE-WGAN structure is used to estimate speech signal and the residual noise by employing a gradient penalty and a weight initialization method for Leaky Rectified Linear Unit (LReLU) and Parametric ReLU (PReLU). The proposed MTAE-WGAN structure with the adopted gradient penalty loss function enhances the speech features and subsequently achieve substantial Phoneme Error Rate (PER) improvements over the stand-alone Deep Denoising Autoencoder (DDAE), MTAE, Redundant Convolutional Encoder-Decoder (R-CED) and Recurrent MTAE (RMTAE) models for robust speech recognition.

• Journal of Korea Game Society
• /
• v.18 no.2
• /
• pp.89-98
• /
• 2018

The recent Neural Network technique has shown good performance in content generation such as image generation in addition to the conventional classification problem and clustering problem solving. In this study, we propose an image generation method using artificial neural network as a next generation content creation technique. The proposed artificial neural network model receives two images and combines them into a new image by taking color from one image and shape from the other image. This model is made up of Convolutional Neural Network, which has two encoders for extracting color and shape from images, and a decoder for taking all the values of each encoder and generating a combination image. The result of this work can be applied to various 2D image generation and modification works in game development process at low cost.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.7C
• /
• pp.611-617
• /
• 2010

In this paper, we investigate the performances of the serial concatenated convolutional codes (SCCC) and low-density parity-check (LDPC) codes on perpendicular magnetic recording (PMR) channels. We discuss the performance of two systems when user bit-densities are 1.7, 2.0, 2.4 and 2.8, respectively. The SCCC system is less complex than LDPC system. The SCCC system consists of recursive systematic convolutional (RSC) codes encoder/decoder, precoder and random interleaver. The decoding algorithm of the SCCC system is the soft message-passing algorithm and the decoding algorithm of the LDPC system is the log domain sum-product algorithm (SPA). When we apply the iterative decoding between channel detector and the error control codes (ECC) decoder, the SCCC system is compatible with the LDPC system even at the high user bit density.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.11a
• /
• pp.180-181
• /
• 2019

This paper proposes a rate control algorithm for intra coding frame in HEVC encoder using a deep learning approach. The proposed algorithm is designed for CTU level bit allocation in intra frame by considering visual features spatially and temporally. Our features are generated using visual geometry group (VGG-16) with deep convolutional layers, then it is used for bit allocation per each CTU within an intra frame. According to our experiments, the proposed algorithm can achieve -2.04% Luma component BD-rate gain with minimal bit accuracy loss against the HM-16.20 rate control model.

• Proceedings of the IEEK Conference
• /
• 2000.07a
• /
• pp.140-143
• /
• 2000

In this paper, forward error control coding in multicarrier direct sequence code division multiple access (DS/CDMA) systems is considered. In order to accommodate a number of coding rates easily and make the encoder and do-coder structure simple, we use the rate compatible punctured convolutional (RCPC) code. We obtain data throughputs at several coding rates and choose the coding rate which has the highest data throughput in the SINR sense. To achieve maximum data throughput, a rate adaptive system using channel state information (the SINR estimate) is proposed. The SINR estimate is obtain by the soft decision Viterbi decoding metric. We show that the proposed rate adaptive convolutionally coded multicarrier DS/CDMA system can enhance spectral efficiency and provide frequency diversity.

• Journal of Information Processing Systems
• /
• v.15 no.5
• /
• pp.1171-1178
• /
• 2019

This study proposes a deep neural network model based on an encoder-decoder structure for visual dialogs. Ongoing linguistic understanding of the dialog history and context is important to generate correct answers to questions in visual dialogs followed by questions and answers regarding images. Nevertheless, in many cases, a visual understanding that can identify scenes or object attributes contained in images is beneficial. Hence, in the proposed model, by employing a separate person detector and an attribute recognizer in addition to visual features extracted from the entire input image at the encoding stage using a convolutional neural network, we emphasize attributes, such as gender, age, and dress concept of the people in the corresponding image and use them to generate answers. The results of the experiments conducted using VisDial v0.9, a large benchmark dataset, confirmed that the proposed model performed well.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.11a
• /
• pp.81-83
• /
• 2018

Fully Convolutional Network(FCN)은 기존의 방법보다 뛰어난 성능을 보였지만, FCN은 RGB 정보만을 사용하기 때문에 세밀한 예측이 필요한 장면에서는 다소 부족한 성능을 보였다. 이를 해결하기 위해 인코더-디코더 구조를 이용하여 RGB와 깊이의 멀티 모달을 활용하기 위한 FuseNet이 제안되었다. 하지만, FuseNet에서는 RGB와 깊이 브랜치 사이의 융합은 있지만, 인코더와 디코더 사이의 특징 지도를 융합하지 않는다. 본 논문에서는 FCN의 디코더 부분의 업샘플링 과정에서 이전 계층의 결과와 2배 업샘플링한 결과를 융합하는 스킵 레이어를 적용하여 FuseNet의 모달리티를 잘 활용하여 성능을 개선했다. 본 실험에서는 NYUDv2와 SUNRGBD 데이터 셋을 사용했으며, 전체 정확도는 각각 77%, 65%이고, 평균 IoU는 47.4%, 26.9%, 평균 정확도는 67.7%, 41%의 성능을 보였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.648-651
• /
• 2023

Video anomaly detection aims to detect abnormal events. Motivated by the power of transformers recently shown in vision tasks, we propose a novel transformer-based network for video anomaly detection. To capture long-range information in video, we employ a multi-scale transformer as an encoder. A convolutional decoder is utilized to predict the future frame from the extracted multi-scale feature maps. The proposed method is evaluated on three benchmark datasets: USCD Ped2, CUHK Avenue, and ShanghaiTech. The results show that the proposed method achieves better performance compared to recent methods.

• Tunnel and Underground Space
• /
• v.29 no.3
• /
• pp.184-196
• /
• 2019

This study aims to extract a 3D image of micro-cracks generated by hydraulic fracturing tests, using the deep learning method and X-ray computed tomography images. The pixel-level cracks are difficult to be detected via conventional image processing methods, such as global thresholding, canny edge detection, and the region growing method. Thus, the convolutional neural network-based encoder-decoder network is adapted to extract and analyze the micro-crack quantitatively. The number of training data can be acquired by dividing, rotating, and flipping images and the optimum combination for the image augmentation method is verified. Application of the optimal image augmentation method shows enhanced performance for not only the validation dataset but also the test dataset. In addition, the influence of the original number of training data to the performance of the deep learning-based neural network is confirmed, and it leads to succeed the pixel-level crack detection.

• Journal of Korea Multimedia Society
• /
• v.24 no.8
• /
• pp.979-987
• /
• 2021

Multiple Sclerosis (MS) can be early diagnosed by detecting lesions in brain magnetic resonance images (MRI). Unsupervised anomaly detection methods based on autoencoder have been recently proposed for automated detection of MS lesions. However, these autoencoder-based methods were developed only for 2D images (e.g. 2D cross-sectional slices) of MRI, so do not utilize the full 3D information of MRI. In this paper, therefore, we propose a novel 3D autoencoder-based framework for detection of the lesion volume of MS in MRI. We first define a 3D convolutional neural network (CNN) for full MRI volumes, and build each encoder and decoder layer of the 3D autoencoder based on 3D CNN. We also add a skip connection between the encoder and decoder layer for effective data reconstruction. In the experimental results, we compare the 3D autoencoder-based method with the 2D autoencoder models using the training datasets of 80 healthy subjects from the Human Connectome Project (HCP) and the testing datasets of 25 MS patients from the Longitudinal multiple sclerosis lesion segmentation challenge, and show that the proposed method achieves superior performance in prediction of MS lesion by up to 15%.