• 대한원격탐사학회지
• /
• 제38권6_1호
• /
• pp.1505-1514
• /
• 2022

영상정합은 다시기 및 다중센서 고해상도 위성영상을 효과적으로 활용하기 위해 필수적으로 선행되는 중요한 과정이다. 널리 각광받고 있는 딥러닝 기법은 위성영상에서 복잡하고 세밀한 특징을 추출하여 영상 간 빠르고 정확한 유사도 판별에 사용될 수 있음에도 불구하고, 학습자료의 양과 질이 결과에 영향을 미치는 딥러닝 모델의 한계와 고해상도 위성영상 기반 학습자료 구축의 어려움에 따라 고해상도 위성영상의 정합에는 제한적으로 적용되어 왔다. 이에 본 연구는 영상정합에서 가장 많은 시간을 소요하는 정합쌍 추출 과정에서 딥러닝 기반 기법의 적용성을 확인하기 위하여, 편향성이 존재하는 고해상도 위성영상 데이터베이스로부터 딥러닝 영상매칭 학습자료를 구축하고 학습자료의 구성이 정합쌍 추출 정확도에 미치는 영향을 분석하였다. 학습자료는 12장의 다시기 및 다중센서 고해상도 위성영상에 대하여 격자 기반의 Scale Invariant Feature Transform(SIFT) 알고리즘을 이용하여 추출한 영상쌍에 참과 거짓의 레이블(label)을 할당한 정합쌍과 오정합쌍의 집합으로 구축되도록 하였다. 구축된 학습자료로부터 정합쌍 추출을 위해 제안된 Siamese convolutional neural network (SCNN) 모델은 동일한 두 개의 합성곱 신경망 구조에 한 쌍을 이루는 두 영상을 하나씩 통과시킴으로써 학습을 진행하고 추출된 특징의 비교를 통해 유사도를 판별한다. 본 연구를 통해 고해상도 위성영상 데이터 베이스로부터 취득된 자료를 딥러닝 학습자료로 활용 가능하며 이종센서 영상을 적절히 조합하여 영상매칭 과정의 효율을 높일 수 있음을 확인하였다. 다중센서 고해상도 위성영상을 활용한 딥러닝 기반 영상매칭 기법은 안정적인 성능을 바탕으로 기존 수작업 기반의 특징 추출 방법을 대체하고, 나아가 통합적인 딥러닝 기반 영상정합 프레임워크로 발전될 것으로 기대한다.

• 정보보호학회논문지
• /
• 제24권6호
• /
• pp.1139-1145
• /
• 2014

지능형 서비스를 제공하는 로봇에서 특정 사람을 인지하거나 구별하는 인식 기술은 매우 중요하다. 기존 단일 거리 얼굴 영상을 학습으로 사용한 얼굴 인식 알고리즘은 원거리로 갈수록 얼굴 인식률이 떨어지는 문제점이 있다. 실제 거리별 얼굴 영상을 이용한 방법은 얼굴 인식률은 향상되지만, 사용자 협조가 요구되는 단점이 있다. 본 논문에서는 줌카메라를 통해 거리별 얼굴 영상을 획득하여 학습으로 사용하는 LDA 기반 원거리 얼굴 인식을 제안한다. 제안하는 방법은 기존 단일거리 얼굴 영상을 학습으로 이용한 방법에 비해 7.8% 향상된 성능을 보였고, 거리별 얼굴 영상을 학습으로 이용한 방법과 비교했을 때 8.0% 저하된 성능을 보였다. 그러나 거리별 얼굴 영상을 취득하기 위해 추가적인 시간과 사용자 협조가 요구되지 않는 장점이 있다.

• 한국치위생학회지
• /
• 제7권4호
• /
• pp.365-379
• /
• 2007

The objective of the current study is to provide basic data necessary for the development of systematic program that is required for the systematic occupational work training of dental hygienists who newly employed at dental hospitals or clinics. The results of the surveys were listed as in below. The survey was conducted for 175 experienced dental hygienists who are in charge of occupational work training in 6 regions(Seoul, Kyunggi, Busan, Ulsan, Kwangju, Chungnam, Kyungnam) of the country where the occupational work training for new dental hygienists is systematically operated. 1. The recognition of experienced dental hygienists for the importance of occupational work training revealed that image training was the most importantly recognized by dental hygienists in Seoul Kyunggi regions(pE.01). In case of Busan region, periodontic training and conservative dentistry training were the most importantly recognized, and customer service training was mostly highly recognized in Ulsan region(pE.01). In case of Kwangjuregion, dental health insurance claim training was recognized as most important subject, and Patient consultation training was the most importantly recognized in Chungnam region. In case of Kyungnam region. Oral surgery was recognized as the most important training subject. 2. Regard on the importance of the range of occupational work training, the experienced dental hygienists with less than 2 years of experience were found to recognize the training of greeting and naming most importantly, the dental hygienists with 2~3 years of experience most importantly recognized oral surgery, and the dental hygienists with 4~5 year of experience were found to recognize conservative training most importantly. In case of dental hygienists having 6~9 year of experience recognized periodontic and conservation trainings as the most important subjects, and the dental hygienist having more than 10 years of experience were found to recognize conservative and image trainings mostly importantly.

• 융복합기술연구소 논문집
• /
• 제11권1호
• /
• pp.1-5
• /
• 2021

Deep learning based methods achieve state-of-the-art accuracy, however, they typically rely on supervised training with large labeled datasets. It is known in many medical applications that labeling medical images requires significant expertise and much time, and typical hand-tuned approaches for data augmentation fail to capture the complex variations in such images. This paper proposes a 3D image augmentation method to overcome these difficulties. It allows us to enrich diversity of training data samples that is essential in medical image segmentation tasks, thus reducing the data overfitting problem caused by the fact the scale of medical image dataset is typically smaller. Our numerical experiments demonstrate that the proposed approach provides significant improvements over state-of-the-art methods for 3D medical image segmentation.

• 한국인터넷방송통신학회논문지
• /
• 제21권6호
• /
• pp.111-116
• /
• 2021

초분광 영상은 적외선 영역의 전자기파 대역을 수백 개의 파장으로 나누어 영상화한 데이터로 다양한 분야에서 물체를 찾거나 분류하는 것에 활용된다. 최근에는 딥러닝을 사용하여 분류하는 방법이 주목받고 있지만 초분광 영상 데이터의 특성으로 인해 초분광 영상을 학습 데이터로 사용하기 위해서는 기존의 가시광 영상과는 다른 처리 기법이 필요하다. 이를 위해 초분광 큐브에서 특정 파장의 영상을 선택하여 Ground Truth 작업을 수행하고 환경정보를 포함하여 데이터를 관리하는 소프트웨어를 개발하였다. 본 논문에서는 해당 소프트웨어의 구성과 기능에 대하여 설명한다.

• 한국군사과학기술학회지
• /
• 제22권1호
• /
• pp.49-59
• /
• 2019

Generative adversarial networks(GANs) have received great attention in the machine learning field for their capacity to model high-dimensional and complex data distribution implicitly and generate new data samples from the model distribution. This paper investigates the model training methodology, architecture, and various applications of generative adversarial networks. Experimental evaluation is also conducted for generating synthetic image dataset for defense using two types of GANs. The first one is for military image generation utilizing the deep convolutional generative adversarial networks(DCGAN). The other is for visible-to-infrared image translation utilizing the cycle-consistent generative adversarial networks(CycleGAN). Each model can yield a great diversity of high-fidelity synthetic images compared to training ones. This result opens up the possibility of using inexpensive synthetic images for training neural networks while avoiding the enormous expense of collecting large amounts of hand-annotated real dataset.

• 대한의용생체공학회:의공학회지
• /
• 제44권4호
• /
• pp.293-301
• /
• 2023

Scoliosis is a three-dimensional deformation of the spine that is a deformity induced by physical or disease-related causes as the spine is rotated abnormally. Early detection has a significant influence on the possibility of nonsurgical treatment. To train a deep learning model with preprocessed images and to evaluate the results with and without data augmentation to enable the diagnosis of scoliosis based only on a chest X-ray image. The preprocessed images in which only the spine, rib contours, and some hard tissues were left from the original chest image, were used for learning along with the original images, and three CNN(Convolutional Neural Networks) models (VGG16, ResNet152, and EfficientNet) were selected to proceed with training. The results obtained by training with the preprocessed images showed a superior accuracy to those obtained by training with the original image. When the scoliosis image was added through data augmentation, the accuracy was further improved, ultimately achieving a classification accuracy of 93.56% with the ResNet152 model using test data. Through supplementation with future research, the method proposed herein is expected to allow the early diagnosis of scoliosis as well as cost reduction by reducing the burden of additional radiographic imaging for disease detection.

• International journal of advanced smart convergence
• /
• 제12권2호
• /
• pp.25-33
• /
• 2023

The purpose of this study is to evaluate the impact of intermediate features on FER performance. To achieve this objective, intermediate features were extracted from the input images at specific layers (FM1~FM4) of the pre-trained network (Resnet-18). These extracted intermediate features and original images were used as inputs to the vision transformer (ViT), and the FER performance was compared. As a result, when using a single image as input, using intermediate features extracted from FM2 yielded the best performance (training accuracy: 94.35%, testing accuracy: 75.51%). When using the original image as input, the training accuracy was 91.32% and the testing accuracy was 74.68%. However, when combining the original image with intermediate features as input, the best FER performance was achieved by combining the original image with FM2, FM3, and FM4 (training accuracy: 97.88%, testing accuracy: 79.21%). These results imply that incorporating intermediate features alongside the original image can lead to superior performance. The findings can be referenced and utilized when designing the preprocessing stages of a deep learning model in FER. By considering the effectiveness of using intermediate features, practitioners can make informed decisions to enhance the performance of FER systems.

• 대한원격탐사학회지
• /
• 제38권1호
• /
• pp.103-110
• /
• 2022

Since satellite images generally include clouds in the atmosphere, it is essential to detect or mask clouds before satellite image processing. Clouds were detected using physical characteristics of clouds in previous research. Cloud detection methods using deep learning techniques such as CNN or the modified U-Net in image segmentation field have been studied recently. Since image segmentation is the process of assigning a label to every pixel in an image, precise pixel-based dataset is required for cloud detection. Obtaining accurate training datasets is more important than a network configuration in image segmentation for cloud detection. Existing deep learning techniques used different training datasets. And test datasets were extracted from intra-dataset which were acquired by same sensor and procedure as training dataset. Different datasets make it difficult to determine which network shows a better overall performance. To verify the effectiveness of the cloud detection network such as Cloud-Net, two types of networks were trained using the cloud dataset from KOMPSAT-3 images provided by the AIHUB site and the L8-Cloud dataset from Landsat8 images which was publicly opened by a Cloud-Net author. Test data from intra-dataset of KOMPSAT-3 cloud dataset were used for validating the network. The simulation results show that the network trained with KOMPSAT-3 cloud dataset shows good performance on the network trained with L8-Cloud dataset. Because Landsat8 and KOMPSAT-3 satellite images have different GSDs, making it difficult to achieve good results from cross-sensor validation. The network could be superior for intra-dataset, but it could be inferior for cross-sensor data. It is necessary to study techniques that show good results in cross-senor validation dataset in the future.

• 한국정보전자통신기술학회논문지
• /
• 제15권5호
• /
• pp.380-386
• /
• 2022

본 논문은 적대적 생성 신경망 기법의 하나인 Pix2Pix를 활용하여 컬러색상을 입히는 경우 학습된 이미지의 빛 반사 정도에 따라 예측결과가 손상되어 나오는 부분에 집중하여 Pix2Pix 모델 적용 전 이미지 전처리 프로세스 및 모델 최적화를 위한 파라미터 테이블을 구성한다. 기존 논문에 나온 Pix2Pix 모델을 활용하여 실생활에 적용하기 위해서는 해상도에 대한 확장성을 고려해야한다. 학습 및 예측결과 이미지 해상도를 키우기 위해서는 동시에 모델의 커널 사이즈 등을 같이 맞춰주는 부분을 수정해줘야 하는데 이부분은 파라미터로 튜닝 가능하도록 설계했다. 또한 본 논문에서는 예측결과가 빛 반사에 의해 손상된 부분만 별도 처리하는 로직을 같이 구성하여 예측결과를 왜곡시키지 않는 전처리 로직을 구성하였다. 따라서 활용성을 개선하기 위하여 Pix2Pix 모델의 학습이미지에 공통적인 빛반사 튜닝 필터를 적용하는 부분과 파라미터 구성부분을 추가하여 모델 정확도를 개선하였다.