• 로봇학회논문지
• /
• 제14권1호
• /
• pp.8-13
• /
• 2019

Visual place recognition is widely researched area in robotics, as it is one of the elemental requirements for autonomous navigation, simultaneous localization and mapping for mobile robots. However, place recognition in changing environment is a challenging problem since a same place look different according to the time, weather, and seasons. This paper presents a feature extraction method using a deep convolutional auto-encoder to recognize places under severe appearance changes. Given database and query image sequences from different environments, the convolutional auto-encoder is trained to predict the images of the desired environment. The training process is performed by minimizing the loss function between the predicted image and the desired image. After finishing the training process, the encoding part of the structure transforms an input image to a low dimensional latent representation, and it can be used as a condition-invariant feature for recognizing places in changing environment. Experiments were conducted to prove the effective of the proposed method, and the results showed that our method outperformed than existing methods.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송∙미디어공학회 2022년도 하계학술대회
• /
• pp.1286-1289
• /
• 2022

최근 영상 군집화 분야는 딥러닝 모델에게 Self-supervision을 주거나 unlabeled 영상에 유사-레이블을 주는 방식으로 연구되고 있다. 또한, 고차원 컬러 자연 영상에 대해 잘 압축된 특징 벡터를 추출하는 것은 군집화에 있어 중요한 기준이 된다. 본 연구에서는 자연 영상에 대한 Convolutional Auto Encoder의 특징 추출 성능을 평가하기 위해 설계한 실험 방법을 소개한다. 특히 모델의 특징 추출 능력을 순수하게 확인하기 위하여 Self-supervision 및 유사-레이블을 제공하지 않은 채 Naive한 모델의 결과를 분석할 것이다. 먼저 실험을 위해 설계된 4가지 비지도학습 모델의 복원 결과를 통해 모델별 학습 정도를 확인한다. 그리고 비지도 모델이 다량의 unlabeled 영상으로 학습되어도 더 적은 labeled 데이터로 학습된 지도학습 모델의 특징 추출 성능에 못 미침을 특징 벡터의 군집화 및 분류 실험 결과를 통해 확인한다. 또한, 지도학습 모델에 데이터셋 간 교차 학습을 수행하여 출력된 특징 벡터의 군집화 및 분류 성능도 확인한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권9호
• /
• pp.3823-3840
• /
• 2020

Recently, auto-encoder has emerged as the most popular method in convolutional neural network (CNN) based image compression and has achieved impressive performance. In the traditional auto-encoder based image compression model, the encoder simply sends the features of last layer to the decoder, which cannot allocate bits over different spatial regions in an efficient way. Besides, these methods do not fully exploit the contextual information under different receptive fields for better reconstruction performance. In this paper, to solve these issues, a novel auto-encoder model is designed for image compression, which can effectively transmit the hierarchical features of the encoder to the decoder. Specifically, we first propose an adaptive bit-allocation strategy, which can adaptively select an importance channel. Then, we conduct the multiply operation on the generated importance mask and the features of the last layer in our proposed encoder to achieve efficient bit allocation. Moreover, we present an additional novel perceptual loss function for more accurate image details. Extensive experiments demonstrated that the proposed model can achieve significant superiority compared with JPEG and JPEG2000 both in both subjective and objective quality. Besides, our model shows better performance than the state-of-the-art convolutional neural network (CNN)-based image compression methods in terms of PSNR.

• 한국멀티미디어학회논문지
• /
• 제23권1호
• /
• pp.1-7
• /
• 2020

Clustering methods essentially take a two-step approach; extracting feature vectors for dimensionality reduction and then employing clustering algorithm on the extracted feature vectors. However, for clustering images, the traditional clustering methods such as stacked auto-encoder based k-means are not effective since they tend to ignore the local information. In this paper, we propose a method first to effectively reduce data dimensionality using convolutional auto-encoder to capture and reflect the local information and then to accurately cluster similar data samples by using a hierarchical clustering approach. The experimental results confirm that the clustering results are improved by using the proposed model in terms of clustering accuracy and normalized mutual information.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권4호
• /
• pp.1689-1703
• /
• 2020

When data is transmitted over an unreliable channel, the error of the data packet may result in serious degradation. The multiple description coding (MDC) can solve this problem and save transmission costs. In this paper, we propose a deep multiple description coding network (MDCN) to realize efficient image compression. Firstly, our network framework is based on convolutional auto-encoder (CAE), which include multiple description encoder network (MDEN) and multiple description decoder network (MDDN). Secondly, in order to obtain high-quality reconstructed images at low bit rates, the encoding network and decoding network are integrated into an end-to-end compression framework. Thirdly, the multiple description decoder network includes side decoder network and central decoder network. When the decoder receives only one of the two multiple description code streams, side decoder network is used to obtain side reconstructed image of acceptable quality. When two descriptions are received, the high quality reconstructed image is obtained. In addition, instead of quantization with additive uniform noise, and SSIM loss and distance loss combine to train multiple description encoder networks to ensure that they can share structural information. Experimental results show that the proposed framework performs better than traditional multiple description coding methods.

• 융합정보논문지
• /
• 제12권1호
• /
• pp.31-38
• /
• 2022

본 논문은 전문가 독립적 비지도 신경망 학습 기반 다변량 시계열 데이터 분석 모델인 MSCRED(Multi-Scale Convolutional Recurrent Encoder-Decoder)의 실제 현장에서의 적용과 Auto-encoder 기반인 MSCRED 모델의 한계인, 학습 데이터가 오염되지 않아야 된다는 점을 극복하기 위한 학습 데이터 샘플링 기법인 Subset Sampling Validation을 제시한다. 라벨 분류가 되어있는 발전소 장비의 진동 데이터를 이용하여 1) 학습 데이터에 비정상 데이터가 섞여 있는 상황을 재현하고, 이를 학습한 경우 2) 1과 같은 상황에서 Subset Sampling Validation 기법을 통해 학습 데이터에서 비정상 데이터를 제거한 경우의 Anomaly Score를 비교하여 MSCRED와 Subset Sampling Validation 기법을 유효성을 평가한다. 이를 통해 본 논문은 전문가 독립적이며 오류 데이터에 강한 이상 진단 프레임워크를 제시해, 다양한 다변량 시계열 데이터 분야에서의 간결하고 정확한 해결 방법을 제시한다.

• Nuclear Engineering and Technology
• /
• 제53권3호
• /
• pp.888-893
• /
• 2021

The International Atomic Energy Agency has developed a tomographic imaging system for accomplishing the total fuel rod-by-rod verification time of fuel assemblies within the order of 1-2 h, however, there are still limitations for some fuel types. The aim of this study is to develop a deep learning-based denoising process resulting in increasing the tomographic image acquisition speed of fuel assembly compared to the conventional techniques. Convolutional AutoEncoder (CAE) was employed for denoising the low-quality images reconstructed by filtered back-projection (FBP) algorithm. The image data set was constructed by the Monte Carlo method with the FBP and ground truth (GT) images for 511 patterns of missing fuel rods. The de-noising performance of the CAE model was evaluated by comparing the pixel-by-pixel subtracted images between the GT and FBP images and the GT and CAE images; the average differences of the pixel values for the sample image 1, 2, and 3 were 7.7%, 28.0% and 44.7% for the FBP images, and 0.5%, 1.4% and 1.9% for the predicted image, respectively. Even for the FBP images not discriminable the source patterns, the CAE model could successfully estimate the patterns similarly with the GT image.

• 한국컴퓨터정보학회논문지
• /
• 제28권3호
• /
• pp.25-33
• /
• 2023

CT 영상의 획득 및 전송 등의 과정에서 발생하는 잡음은 영상의 질을 저하시키는 요소로 작용한다. 따라서 이를 해결하기 위한 잡음제거는 영상처리에서 중요한 전처리 과정이다. 본 논문에서는 딥러닝의 convolutional autoencoder (CAE) 모형에서 기존 컨볼루션 연산 대신 deformable 컨볼루션 연산을 적용한 deformable convolutional autoencoder (DeCAE) 모형을 이용하여 잡음을 제거하고자 한다. 여기서 deformable 컨볼루션 연산은 기존 컨볼루션 연산보다 유연한 영역에서 영상의 특징들을 추출할 수 있다. 제안된 DeCAE 모형은 기존 CAE 모형과 같은 인코더-디코더 구조로 되어있으나 효율적인 잡음제거를 위해 인코더는 deformable 컨볼루션 층으로 구성하고, 디코더는 기존 컨볼루션 층으로 구성하였다. 본 논문에서 제안된 DeCAE 모형의 성능 평가를 위해 다양한 잡음, 즉, 가우시안 잡음, 임펄스 잡음 그리고 포아송 잡음에 의해 훼손된 CT 영상을 대상으로 실험하였다. 성능 실험 결과, DeCAE 모형은 전통적인 필터 즉, Mean 필터, Median 필터와 이를 개선한 Bilateral 필터, NL-means 방법 뿐만 아니라 기존의 CAE 모형보다 정성적이고, 정량적인 척도 즉, MAE (Mean Absolute Error), PSNR (Peak Signal-to-Noise Ratio) 그리고 SSIM (Structural Similarity Index Measure) 면에서 우수한 결과를 보였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권12호
• /
• pp.4909-4926
• /
• 2020

Cyberattacks penetrate the server and perform various malicious acts such as stealing confidential information, destroying systems, and exposing personal information. To achieve this, attackers perform various malicious actions by infecting endpoints and accessing the internal network. However, the current countermeasures are only anti-viruses that operate in a signature or pattern manner, allowing initial unknown attacks. Endpoint Detection and Response (EDR) technology is focused on providing visibility, and strong countermeasures are lacking. If you fail to respond to the initial attack, it is difficult to respond additionally because malicious behavior like Advanced Persistent Threat (APT) attack does not occur immediately, but occurs over a long period of time. In this paper, we propose a technique that detects an unknown attack using an event log without prior knowledge, although the initial response failed with anti-virus. The proposed technology uses a combination of AutoEncoder and 1D CNN (1-Dimention Convolutional Neural Network) based on semi-supervised learning. The experiment trained a dataset collected over a month in a real-world commercial endpoint environment, and tested the data collected over the next month. As a result of the experiment, 37 unknown attacks were detected in the event log collected for one month in the actual commercial endpoint environment, and 26 of them were verified as malicious through VirusTotal (VT). In the future, it is expected that the proposed model will be applied to EDR technology to form a secure endpoint environment and reduce time and labor costs to effectively detect unknown attacks.

• 한국음향학회지
• /
• 제38권6호
• /
• pp.670-677
• /
• 2019

음성 또는 음향 이벤트 신호에서 발생하는 배경 잡음은 인식기의 성능을 저하시키는 원인이 되며, 잡음에 강인한 특징을 찾는데 많은 노력을 필요로 한다. 본 논문에서는 딥러닝을 기반으로 다중작업 오토인코더(Multi-Task AutoEncoder, MTAE) 와 와설스타인식 생성적 적대 신경망(Wasserstein GAN, WGAN)의 장점을 결합하여, 잡음이 섞인 음향신호에서 잡음과 음성신호를 추정하는 네트워크를 제안한다. 본 논문에서 제안하는 MTAE-WGAN는 구조는 구배 페널티(Gradient Penalty) 및 누설 Leaky Rectified Linear Unit (LReLU) 모수 Parametric ReLU (PReLU)를 활용한 변수 초기화 작업을 통해 음성과 잡음 성분을 추정한다. 직교 구배 페널티와 파라미터 초기화 방법이 적용된 MTAE-WGAN 구조를 통해 잡음에 강인한 음성특징 생성 및 기존 방법 대비 음소 오인식률(Phoneme Error Rate, PER)이 크게 감소하는 성능을 보여준다.