• Korean Journal of Remote Sensing
• /
• v.24 no.4
• /
• pp.299-307
• /
• 2008

It is possible to extract qualified ground control points (GCPs) from SAR data itself without published maps. TerraSAR-X data that are one of highest spatial resolution among civilian SAR systems is now available. In this study, a sophisticated method for GCP extraction from TerraSAR-X data was tested and the quality of the extracted GCPs was evaluated. Mean values of the distance errors were 0.11m and -3.96 m with standard deviations of 6.52 m and 5.11 m in easting and northing, respectively. The result is one of the best among GCPs possibly extracted from any civilian remote sensing systems. The extracted GCPs were used for geo-rectification of IKONOS image. The method used in this study can be applied to KOMPSAT-5 for geo-rectification of high-resolution optic images acquired by KOMPSAT-2 or follow-up missions.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.37 no.5
• /
• pp.379-388
• /
• 2019

Recently, a technique for acquiring spatial information data using UAV (Unmanned Aerial Vehicle) has been greatly developed. It is a very crucial issue of the GIS (Geographic Information System) mapping system that passes way point in the unmanned airframe and finally measures the accurate image and stable localization to the desired destination. Though positioning using DGPS (Differential Global Navigation System) or RTK-GPS (Real Time Kinematic-GPS) guarantee highly accurate, they are more expensive than the construction of a single positioning system using a single GPS. In the case of a low-priced single GPS system, the stability of the positioning data deteriorates. Therefore, it is necessary to supplement the uncertainty of the absolute position data of the UAV and to improve the accuracy of the current position data economically in the operating state of the UAV. The aim of this study was to present an algorithm enhancing the stability of position data in a single GPS mode of UAV with multiple GPS. First, the arrangement of multiple GPS receivers through the center of gravity of the UAV were examined. Next, MD (Mahalanobis Distance) is applied to detect instantaneous errors of GPS data in advance and eliminate outliers to increase the accuracy of previously collected multiple GPS data. Processing procedure for multiple GPS reception data by applying the center of the triangular method were presented to improve the position accuracy. Second, UAV navigation systems integrated multiple GPS through configuration of the UAV specifications were implemented. Using the unmanned airframe equipped with multiple GPS receivers, GPS data is measured with the TCM (Triangular Center Method). In addition, UAV equipped with multiple GPS were operated in study area and locational accuracy of multiple GPS of UAV with VRS (Virtual Reference Station) GNSS surveying were compared. The result showed that the error factors are compensated, and the error range are reduced, resulting in the reliability of the corrected value. In conclusion, the result in this paper is expected to realize high-precision position estimation at low cost in UAV using multiple low-cost GPS receivers.

• Imaging Science in Dentistry
• /
• v.51 no.3
• /
• pp.299-306
• /
• 2021

Purpose: This study aimed to propose a fully automatic landmark identification model based on a deep learning algorithm using real clinical data and to verify its accuracy considering inter-examiner variability. Materials and Methods: In total, 950 lateral cephalometric images from Yonsei Dental Hospital were used. Two calibrated examiners manually identified the 13 most important landmarks to set as references. The proposed deep learning model has a 2-step structure-a region of interest machine and a detection machine-each consisting of 8 convolution layers, 5 pooling layers, and 2 fully connected layers. The distance errors of detection between 2 examiners were used as a clinically acceptable range for performance evaluation. Results: The 13 landmarks were automatically detected using the proposed model. Inter-examiner agreement for all landmarks indicated excellent reliability based on the 95% confidence interval. The average clinically acceptable range for all 13 landmarks was 1.24 mm. The mean radial error between the reference values assigned by 1 expert and the proposed model was 1.84 mm, exhibiting a successful detection rate of 36.1%. The A-point, the incisal tip of the maxillary and mandibular incisors, and ANS showed lower mean radial error than the calibrated expert variability. Conclusion: This experiment demonstrated that the proposed deep learning model can perform fully automatic identification of cephalometric landmarks and achieve better results than examiners for some landmarks. It is meaningful to consider between-examiner variability for clinical applicability when evaluating the performance of deep learning methods in cephalometric landmark identification.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.6
• /
• pp.733-739
• /
• 2019

Water level measurement is highly demanding in IoT monitoring areas such as smart factory, smart farm, and smart fish farm. However, existing water level indicators are limited to be used in industrial fields as commercial products due to the high cost of sensors and the complexity of algorithms used. In order to solve these problems, our paper proposed methods using an infrared distance sensor as well as a hall sensor for the water level measurement, both of which are contactless smart sensors. Data errors caused by the inaccuracy of existing sensors were decreased by applying new simple structures so that versatility is enhanced. The performance of our method was validated using experiments based on simulations. We expect that our new water depth indicator can be extended to a general-purpose water level monitoring system based on IoT technology.

• Imaging Science in Dentistry
• /
• v.49 no.2
• /
• pp.159-169
• /
• 2019

Purpose: Cone-beam computed tomography (CBCT) is widely used for 3-dimensional assessments of cranio-maxillo-facial relationships, especially in patients undergoing orthognathic surgery. We have introduced, for reference in CBCT cephalometry, an anatomical mid-sagittal plane (MSP) identified by the nasion, the midpoint between the posterior clinoid processes of the sella turcica, and the basion. The MSP is an updated version of the median plane previously used at our institution for 2D posterior-anterior cephalometry. This study was conducted to test the accuracy of the CBCT measures compared to those obtained using standard posterior-anterior cephalometry. Materials and Methods: Two operators measured the inter-zygomatic distance on 15 CBCT scans using the MSP as a reference plane, and the CBCT measurements were compared with measurements made on patients' posterior-anterior cephalograms. The statistical analysis evaluated the absolute and percentage differences between the 3D and 2D measurements. Results: As demonstrated by the absolute mean difference (roughly 1 mm) and the percentage difference (less than 3%), the MSP showed good accuracy on CBCT compared to the 2D plane, especially for measurements of the left side. However, the CBCT measurements showed a high standard deviation, indicating major variability and low precision. Conclusion: The anatomical MSP can be used as a reliable reference plane for transverse measurements in 3D cephalometry in cases of symmetrical or asymmetrical malocclusion. In patients who suffer from distortions of the skull base, the identification of landmarks might be difficult and the MSP could be unreliable. Becoming familiar with the relevant software could reduce errors and improve reliability.

• Progress in Medical Physics
• /
• v.32 no.3
• /
• pp.59-69
• /
• 2021

Purpose: The relationship between computed tomography (CT) number and electron density (ED) has been investigated in previous studies. However, the role of these measures for guiding cancer treatment remains unclear. Methods: The CT number was plotted against ED for different imaging protocols. The CT number was imported into ED tables for the Pinnacle treatment planning system (TPS) and was used to determine the effect on dose calculations. Conversion tables for radiation dose calculations were generated and subsequently monitored using a dosimeter to determine the effect of different CT scanning protocols and treatment sites. These tables were used to retrospectively recalculate the radiation therapy plans for 41 patients after an incorrect scanning protocol was inadvertently used. The gamma index was further used to assess the dose distribution, percentage dose difference (DD), and distance-to-agreement (DTA). Results: For densities <1.1 g/cm³, the standard deviation of the CT number was ±0.6% and the greatest variation was noted for brain protocol conditions. For densities >1.1 g/cm³, the standard deviation of the CT number was ±21.2% and the greatest variation occurred for the tube voltage and head and neck (H&N) protocol conditions. These findings suggest that the factors most affecting the CT number are the tube voltage and treatment site (brain and H&N). Gamma index analyses for the 41 retrospective clinical cases, as well as brain metastases and H&N tumors, showed gamma passing rates >90% and <90% for the passing criterion of 2%/2 and 1%/1 mm, respectively. Conclusions: The CT protocol should be carefully decided for TPS. The correct protocol should be used for the corresponding TPS based on the treatment site because this especially affects the dose distribution for brain metastases and H&N tumor recognition. Such steps could help reduce systematic errors.

• Korean Journal of Materials Research
• /
• v.31 no.12
• /
• pp.665-671
• /
• 2021

A 1.8 ㎛ thick polycrystalline diamond (PCD) thin film layer is prepared on a Si(100) substrate using hot-filament chemical vapor deposition. Thereafter, its thermal conductivity is measured using the conventional laser flash analysis (LFA) method, a LaserPIT-M2 instrument, and the newly proposed light source thermal analysis (LSTA) method. The LSTA method measures the thermal conductivity of the prepared PCD thin film layer using an ultraviolet (UV) lamp with a wavelength of 395 nm as the heat source and a thermocouple installed at a specific distance. In addition, the microstructure and quality of the prepared PCD thin films are evaluated using an optical microscope, a field emission scanning electron microscope, and a micro-Raman spectroscope. The LFA, LaserPIT-M2, and LSTA determine the thermal conductivities of the PCD thin films, which are 1.7, 1430, and 213.43 W/(m·K), respectively, indicating that the LFA method and LaserPIT-M2 are prone to errors. Considering the grain size of PCD, we conclude that the LSTA method is the most reliable one for determining the thermal conductivity of the fabricated PCD thin film layers. Therefore, the proposed LSTA method presents significant potential for the accurate and reliable measurement of the thermal conductivity of PCD thin films.

• Journal of Korea Society of Industrial Information Systems
• /
• v.27 no.2
• /
• pp.35-44
• /
• 2022

In this paper, we propose an intra-picture prediction method by a quadratic surface modeling method for depth video coding. The pixels of depth video are transformed to 3D coordinates using distance information. A quadratic surface with the smallest error is found by least square method for reference pixels. The reference pixel can be either the upper pixels or the left pixels. In the intra prediction using the quadratic surface, two predcition values are computed for one pixel. Two errors are computed as the square sums of differences between each prediction values and the pixel values of the reference pixels. The pixel sof the block are predicted by the reference pixels and prediction method that they have the lowest error. Comparing with the-state-of-art video coding method, simulation results show that the distortion and the bit rate are improved by up to 5.16% and 5.12%, respectively.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.21 no.2
• /
• pp.17-32
• /
• 2022

This study explored an approach to estimate a flow-density diagram(FD) on a link in highway traffic environment by utilizing probe vehicles' time headway records. To study empirical flow-density diagram(EFD), the probe vehicles with vision sensors were recruited for collecting driving records for nine months and the vision sensor data pre-processing and GIS-based map matching were implemented. Then, we examined the new EFDs to evaluate validity with reference diagrams which is derived from loop detection traffic data. The probability distributions of time headway and distance headway as well as standard deviation of flow and density were utilized in examination. As a result, it turned out that the main factors for estimation errors are the limited number of probe vehicles and bias of flow status. We finally suggest a method to improve the accuracy of EFD model.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.5
• /
• pp.662-667
• /
• 2022

Semantic role labeling task(SRL) is to extract predicate and arguments such as agent, patient, place, time. In the previously SRL task studies, a pipeline method extracting linguistic features of sentence has been proposed, but in this method, errors of each extraction work in the pipeline affect semantic role labeling performance. Therefore, methods using End-to-End neural network model have recently been proposed. In this paper, we propose a neural network model using the Biaffine Average Attention model for SRL task. The proposed model consists of a structure that can focus on the entire sentence information regardless of the distance between the predicate in the sentence and the arguments, instead of LSTM model that uses the surrounding information for prediction of a specific token proposed in the previous studies. For evaluation, we used F1 scores to compare two models based BERT model that proposed in existing studies using F1 scores, and found that 76.21% performance was higher than comparison models.