• Korean Journal of Radiology
• /
• 제2권1호
• /
• pp.42-51
• /
• 2001

Because the tongue is superficially located and the initial manifestation of most diseases occurring there is mucosal change, lingual lesionscan be easily accessed and diagnosed without imaging analysis. Some lingual neoplasms, however, may manifest as a submucosal bulge and be located in a deep portion of the tongue, such as its base; their true characteristics and extent may be recognized only on cross-sectional images such as those obtained by CT or MRI. Some uncommon tongue neoplasms may have characteristic radiologic features, thus permitting quite specific radiologic diagnosis. Lipomas typically manifest at both CT and MR imaging as homogeneous nonenhancing lesions. Relative to subcutaneous fat they are isoattenuating on CT images, and all MR sequences show them as isointense. Due to the paramagnetic properties of melanin, metastases from melanotic melanoma usually demonstrate high signal intensity on T1-weighted MR images and low signal intensity on T2-weighted images. Although the radiologic findings for other submucosal neoplasms are nonspecific, CT and MR imaging can play an important role in the diagnostic work-up of these unusual tumors. Delineation of the extent of the tumor, and recognition and understanding of the spectrum of imaging and the pathologic features of these lesions, often help narrow the differential diagnosis.

• 대한원격탐사학회지
• /
• 제39권4호
• /
• pp.395-407
• /
• 2023

Super-resolution (SR) has great significance in image processing because it enables downstream vision tasks with high spatial resolution. Recently, SR studies have adopted deep learning networks and achieved remarkable SR performance compared to conventional example-based methods. Deep-learning-based SR models generally require low-resolution (LR) images and the corresponding high-resolution (HR) images as training dataset. Due to the difficulties in obtaining real-world LR-HR datasets, most SR models have used only HR images and generated LR images with predefined degradation such as bicubic downsampling. However, SR models trained on simple image degradation do not reflect the properties of the images and often result in deteriorated SR qualities when applied to real-world images. In this study, we propose an image degradation model for HR satellite images based on the modulation transfer function (MTF) of an imaging sensor. Because the proposed method determines the image degradation based on the sensor properties, it is more suitable for training SR models on remote sensing images. Experimental results on HR satellite image datasets demonstrated the effectiveness of applying MTF-based filters to construct a more realistic LR-HR training dataset.

• 한국의류학회지
• /
• 제17권4호
• /
• pp.529-538
• /
• 1993

This study was intended to identify the hierarchy of clothing images, which is expected to be helpful in style classification and product positioning. A questionnaire consisted of 110 words expressing clothing images was developed, and eight clothing photographs were selected as stimuli. 289 female subjects aged between 22 to 37 responded to two of the eight photographs during September, 1991. 110 words were reduced to 62 words based on their independence before conducting factor analysis to identify the constructing factors of clothing images. Nine words with negative connotations were eliminated, because they are not sought in product development. To explain the hierarchy of clothing images, cluster analysis was applied. To observe the association of 53 words, dendrogram was introduced, and to interpret the result, eleven sub clusters were determined. This 11 clusters were continuously combined according to their similarities, until they integrated into one 'clothing image'. Two major division of image clusters were 'graceful and feminine image', and 'mannish and simple image'.

• 한국멀티미디어학회논문지
• /
• 제24권3호
• /
• pp.396-405
• /
• 2021

In surgery to remove pancreatic cancer, it is important to figure out the shape of a patient's pancreas. However, previous studies have a limit to detect a pancreas automatically in abdominal CT images, because the pancreas varies in shape, size and location by patient. Therefore, in this paper, we propose a method of learning various shapes of pancreas according to the patients and adjacent slices using Faster R-CNN based on Inception V2, and automatically detecting the pancreas from abdominal CT images. Model training and testing were performed using the NIH Pancreas-CT Dataset, and intensity normalization was applied to all data to improve pancreatic detection accuracy. Additionally, according to the shape of the pancreas, the test dataset was classified into top, middle, and bottom slices to evaluate the model's performance on each data. The results show that the top data's mAP@.50IoU achieved 91.7% and the bottom data's mAP@.50IoU achieved 95.4%, and the highest performance was the middle data's mAP@.50IoU, 98.5%. Thus, we have confirmed that the model can accurately detect the pancreas in CT images.

• 한국안전학회지
• /
• 제36권1호
• /
• pp.18-25
• /
• 2021

In many developed countries, such as South Korea, efficiently maintaining the aging infrastructures is an important issue. Currently, inspectors visually inspect the infrastructure for maintenance needs, but this method is inefficient due to its high costs, long logistic times, and hazards to the inspectors. Thus, in this paper, a novel crack inspection approach for concrete bridges is proposed using integrated image processing techniques. The proposed approach consists of four steps: (1) training a deep learning model to automatically detect cracks on concrete bridges, (2) acquiring in-situ images using a drone, (3) generating orthomosaic images based on 3D modeling, and (4) detecting cracks on the orthmosaic image using the trained deep learning model. Cascade Mask R-CNN, a state-of-the-art instance segmentation deep learning model, was trained with 3235 crack images that included 2415 hard negative images. We selected the Tancheon overpass, located in Seoul, South Korea, as a testbed for the proposed approach, and we captured images of pier 34-37 and slab 34-36 using a commercial drone. Agisoft Metashape was utilized as a 3D model generation program to generate an orthomosaic of the captured images. We applied the proposed approach to four orthomosaic images that displayed the front, back, left, and right sides of pier 37. Using pixel-level precision referencing visual inspection of the captured images, we evaluated the trained Cascade Mask R-CNN's crack detection performance. At the coping of the front side of pier 37, the model obtained its best precision: 94.34%. It achieved an average precision of 72.93% for the orthomosaics of the four sides of the pier. The test results show that this proposed approach for crack detection can be a suitable alternative to the conventional visual inspection method.

• Korean Journal of Radiology
• /
• 제1권3호
• /
• pp.135-141
• /
• 2000

Objective: To compare the clinical utility of the different imaging techniques used for the evaluation of tracheobronchial diseases. Materials and Methods: Forty-one patients with tracheobronchial diseases [tuberculosis (n = 18), bronchogenic carcinoma (n = 10), congenital abnormality (n = 3), post-operative stenosis (n = 2), and others (n = 8)] underwent chest radiography and spiral CT. Two sets of scan data were obtained: one from routine thick-section axial images and the other from thin-section axial images. Multiplanar reconstruction (MPR) and shaded surface display (SSD) images were obtained from thin-section data. Applying a 5-point scale, two observers compared chest radiography, routine CT, thin-section spiral CT, MPR and SSD imaging with regard to the detection, localization, extent, and characterization of a lesion, information on its relationship with adjacent structures, and overall information. Results: SSD images were the most informative with regard to the detection (3.95±0.31), localization (3.95±0.22) and extent of a lesion (3.85±0.42), and overall information (3.83±0.44), while thin-section spiral CT scans provided most information regarding its relationship with adjacent structures (3.56±0.50) and characterization of the lesion (3.51±0.61). Conclusion: SSD images and thin-section spiral CT scans can provide valuable information for the evaluation of tracheobronchial disease.

• 한국과학교육학회지
• /
• 제28권5호
• /
• pp.453-463
• /
• 2008

이 연구에서는 Draw-An-Environmental-Scientist-Test(DAEST)를 이용하여 학생들의 환경과학자와 환경과학자가 하는 일에 대한 인식과, 환경과학자 그림을 그리는데 영향을 미친 요인을 조사했다. 서울시 소재 일반계 고등학교 3개교에서 1학년과 2학년 학생 413명을 대상으로 DAEST를 실시했다. 연구 결과, 학생들의 환경과학자에 대한 이미지는 과학자의 정형적인 이미지와는 다르게 나타났다. 학생들의 그림에는 환경과학자의 나이와 성의 특징이 뚜렷하게 나타나지 않았다. 또한, 대부분의 학생들은 환경과학자를 야외에서 실험기구와 노트북 컴퓨터 등을 이용하여 환경오염과 관련된 데이터 수집을 하는 사람으로 인식했다. 한편, 학생들은 환경과학자에 대한 그림을 그리는데 영향을 준 요인으로 대중매체, 학교교육, 인터넷 등을 제시했다. 환경과학자와 환경과학자가 하는 일에 대한 인식 및 인식에 영향을 미친 요인은 학생들의 학년과 성에 따라 다르게 나타났다. 이에 대한 교육적 함의를 논의했다.

• 한국실내디자인학회논문집
• /
• 제20권3호
• /
• pp.259-266
• /
• 2011

Today, construction fence screens which were understood simply as structures of separating people from the dangers of construction sites have evolved into landmarks to enhance the images of Cities. Now, construction fence screens should have the function as landmarks stimulating people's sensibility through providing aesthetic values as well as the function of protecting people from the dangers of fields of construction. For this reason, this study tried 1) to show the relationships among these 4 classifications. is 4 realms of Human Beings - Cultural - Ecological - Digital in the expression methods of the construction fence screens of Seoul and 2) to simplify the characteristic elements of construction fence screens by way of extracting with Ecology, Installation Art, Symbolic, Commercial Expression. This study examined whether or not the construction fence screens of Seoul were playing the role as landmarks stimulating people's sensibility with sustainable improvement to the environment of Seoul as public design structures, by investigating some of the construction fence screens of Seoul. Also through analyzing the diversity of expression methods of the construction fence screens of other local governments as well as those of Seoul, we hope that they will be the public structures for enhancing the images of Cities.

• 로봇학회논문지
• /
• 제19권1호
• /
• pp.98-105
• /
• 2024

This paper proposes a new approach with a distance-based regularization to the entropy applied to the NBV (Next-Best-View) selection with NeRF (Neural Radiance Fields). 3D reconstruction requires images from various viewpoints, and selecting where to capture these images is a highly complex problem. In a recent work, image acquisition was derived using NeRF's ray-based uncertainty. While this work was effective for evaluating candidate viewpoints at fixed distances from a camera to an object, it is limited when dealing with a range of candidate viewpoints at various distances, because it tends to favor selecting viewpoints at closer distances. Acquiring images from nearby viewpoints is beneficial for capturing surface details. However, with the limited number of images, its image selection is less overlapped and less frequently observed, so its reconstructed result is sensitive to noise and contains undesired artifacts. We propose a method that incorporates distance-based regularization into entropy, allowing us to acquire images at distances conducive to capturing both surface details without undesired noise and artifacts. Our experiments with synthetic images demonstrated that NeRF models with the proposed distance and entropy-based criteria achieved around 50 percent fewer reconstruction errors than the recent work.

• 대한의용생체공학회:학술대회논문집
• /
• 대한의용생체공학회 1997년도 추계학술대회
• /
• pp.206-209
• /
• 1997

The stereostatic surgery becomes a important part of neurosurgery. The conventional style of stereostatic system uses a input method of film digitizing. It is time consuming and laborious. In this paper, we presented a system that can manage digital images from medical imaging machine to surgery assisting program. CT images were transferred in DICOM format to a surgery assisting computer in a operation room through PACS. The streotatic surgery assisting program processed the digital images and calculated the parameters that were required in steostatic surgery. The assisting program were developed that can detect the reference points automatically, transform CT coordinate to frame coordinate and calculate the RM (Riechert-Mundinger) frame. This system were applied to clinical cases in Seoul National Univ. Hospital. The two advantages of this system were revealed. The processing time from imaging to surgery parameters were much aster than conventional system. The surgery accuracy were more mininute as the digitizing error were reduced. This system were a good application of connecting imaging machines to clinical treatment system through PACS.