• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.7
• /
• pp.1841-1857
• /
• 2023

CenterNet, a novel object detection algorithm without anchor based on key points, regards the object as a single center point for prediction and directly regresses the object's height and width. However, because the objects have different sizes, directly regressing their height and width will make the model difficult to converge and lose the intrinsic relationship between object's width and height, thereby reducing the stability of the model and the consistency of prediction accuracy. For this problem, we proposed an algorithm based on the regression of the diagonal half-length and the center angle, which significantly compresses the solution space of the regression components and enhances the intrinsic relationship between the decoded components. First, encode the object's width and height into the diagonal half-length and the center angle, where the center angle is the angle between the diagonal and the vertical centreline. Secondly, the predicted diagonal half-length and center angle are decoded into two length components. Finally, the position of the object bounding box can be accurately obtained by combining the corresponding center point coordinates. Experiments show that, when using CenterNet as the improved baseline and resnet50 as the Backbone, the improved model achieved 81.6% and 79.7% mAP on the VOC 2007 and 2012 test sets, respectively. When using Hourglass-104 as the Backbone, the improved model achieved 43.3% mAP on the COCO 2017 test sets. Compared with CenterNet, the improved model has a faster convergence rate and significantly improved the stability and prediction accuracy.

• The korean journal of orthodontics
• /
• v.35 no.3 s.110
• /
• pp.216-226
• /
• 2005

The purpose of this study was to test and compare the accuracy and reliability of soft tissue profile predictions generated from two computer software programs (Quick Ceph Image $Pro^{TM}$ (ver 3.0) and $V-Ceph^{TM}$(ver 3.5)) for mandibular set-back surgery. The presurgical and postsurgical lateral cephalograms of 40 patients (20 males and 20 females) were traced on the same acetate paper with the reference taken as the cranial base outline. The presurgical skeletal outlines were digitized onto each computer program and the mandible was moved to mimic the expected surgical procedure with reverence to the mandibular anterior border and lower incisor position of the actual postsurgical skeletal outline. The soft tissue profile was generated and the amount and direction of skeletal movement was calculated with each software. The predicted soft tissue profile was compared to the actual postsurgical soft tissue profile. There were differences between the actual and the predicted surgical soft tissue profile charges in the magnitude and direction, especially the upper lip. lower lip and the soft tissue chin (P<0.05). Quick Ceph had more horizontal measurement errors and thickness errors for the upper lip and lower lip, but V-Ceph had more vertical measurement errors of the lower lip (P<0.05). There was a positive correlation between the prediction errors and the amount of mandibular movements in the vertical position of Sn, the horizontal position of Ls and the upper lip thickness for V-Ceph, and there was a negative correlation in the horizontal position and the thickness of the lower lip for Quick Ceph (P<0.05). However all of the Prediction errors of both imaging softwares were ranged within 3mm, and this was considered to be allowable clinically.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.44 no.3
• /
• pp.229-238
• /
• 2008

This paper describes on the consolidation of AIS and ARPA radar positions by comparing the AIS and ARPA radar information for the tracked ship targets using a PC-based ECDIS in Busan harbor, Korea. The information of AIS and ARPA radar target was acquired independently, and the tracking parameters such as ship's position, COG, SOG, gyro heading, rate of turn, CPA, TCPA, ship s name and MMSI etc. were displayed automatically on the chart of a PC-based ECDIS with radar overlay and ARPA tracking. The ARPA tracking information obtained from the observed radar images of the target ship was compared with the AIS information received from the same vessel to investigate the difference in the position and movement behavior between AIS and ARPA tracked target ships. For the ARPA radar and AIS targets to be consolidated, the differences in range, speed, course, bearing and distance between their targets were estimated to obtain a clear standards for the consolidation of ARPA radar and AIS targets. The average differences between their ranges, their speeds and their courses were 2.06% of the average range, -0.11 knots with the averaged SOG of 11.62 knots, and $0.02^{\circ}$ with the averaged COG of $37.2^{\circ}$, respectively. The average differences between their bearings and between their positions were $-1.29^{\circ}$ and 68.8m, respectively. From these results, we concluded that if the ROT, COG, SOG, and HDG informations are correct, the AIS system can be improved the prediction of a target ship's path and the OOW(Officer of Watch) s ability to anticipate a traffic situation more accurately.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.5
• /
• pp.48-54
• /
• 2018

With the 4th industry era, smart factories are emerging. In the era of multi-product small scale production, unmanned transportation vehicles are rapidly increasing in utilization of unmanned transportation vehicles that carry and arrange goods in the work space. The conventional unmanned vehicle detected its position by using the guided line method and the position based method for indoor location recognition and movement. This method has disadvantages of initial high cost and maintenance / maintenance. In this paper, to solve the disadvantages, the method of predicting the direct path of the unmanned vehicle through the Kalman filter is verified using the white LED marker of the warehouse and the position data and the image data of the white LED marker recognition image. Through this, the reliability of the linear movement which occupies the most part in the lattice structure is secured. It is also expected that the reliance on additional position sensors will also be reduced.

• KIEAE Journal
• /
• v.14 no.1
• /
• pp.67-74
• /
• 2014

There have distinctly been no the installation criteria and maintenance management of BIPV systems, although the BIPV market is consistently going on increasing. In addition, consideration of the BIPV generation quantity which has been installed at several diverse places is currently almost behind within region in Korea. Therefore, the main aim of this study is to evaluate the BIPV generation and to be base data of reducing rate depending on regional installation angles using PVpro which was verified by measured data. Various conditions were an angle of inclination and azimuth under six major cities: Seoul, Daejeon, Daegu, Busan, Gwangju, Jeju-si for the BIPV system generation analysis. As the results, Seoul showed the lowest BIPV generation: 1,054kWh/kWp.year, and Jeju-si have 5percent more generation: 1,108.0kWh/kWp.year than Seoul on horizontal plane. Gwangju and Daejeon turned out to have similar generation of result, and Busan showed the highest generation: 1,193.5kWh/kWp.year, which was increased by over 13percent from Seoul on horizontal plane. Another result, decreasing rate of BIPV generation depending on regional included angle indicate that the best position was located on azimuth: $0^{\circ}$(The south side) following the horizontal position(an angle of inclination: $30^{\circ}$). And the direction on a south vertical position(azimuth: $0^{\circ}$, an angle of inclination: $90^{\circ}$) then turned out reducing rate about 40percent compared with the best one. Therefore, these results would be used to identify the installation angle of the BIPV module as an appropriate position.

• Progress in Medical Physics
• /
• v.29 no.2
• /
• pp.66-72
• /
• 2018

The proper position of a multi-leaf collimator (MLC) is essential for the quality of intensity-modulated radiation therapy (IMRT) and volumetric modulated arc radiotherapy (VMAT) dose delivery. Task Group (TG) 142 provides a quality assurance (QA) procedure for MLC position. Our study investigated the QA validation of the mechanical leaf gap measurement and the maintenance procedure. Two $VitalBeam^{TM}$ systems were evaluated to validate the acceptance of an MLC position. The dosimetric leaf gaps (DLGs) were measured for 6 MV, 6 MVFFF, 10 MV, and 15 MV photon beams. A solid water phantom was irradiated using $10{\times}10cm^2$ field size at source-to-surface distance (SSD) of 90 cm and depth of 10 cm. The portal dose image prediction (PDIP) calculation was implemented on a treatment planning system (TPS) called $Eclipse^{TM}$. A total of 20 VMAT plans were used to confirm the accuracy of dose distribution measured by an electronic portal imaging device (EPID) and those predicted by VMAT plans. The measured leaf gaps were 0.30 mm and 0.35 mm for VitalBeam 1 and 2, respectively. The DLG values decreased by an average of 6.9% and 5.9% after mechanical MLC adjustment. Although the passing rates increased slightly, by 1.5% (relative) and 1.2% (absolute) in arc 1, the average passing rates were still within the good dose delivery level (>95%). Our study shows the existence of a mechanical leaf gap error caused by a degenerated MLC motor. This can be recovered by reinitialization of MLC position on the machine control panel. Consequently, the QA procedure should be performed regularly to protect the MLC system.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.2
• /
• pp.211-218
• /
• 2020

The radio wave propagation characteristics of the transmitter and receiver position change in the indoor environment were predicted through simulation, then the results obtained through the transmission loss measurement were compared and analyzed with the simulation results. The conference room was chosen as the environment for measuring transmission loss, and the radio transmission characteristics of the two environments were compared by selecting the exhibition hall without interior decorations and fixtures. In each indoor environment, the position of the transmitter chose two cases. One located in the center of the front wall and the other in the center of the side wall, and the position of the receiver moved along the centerline of the conference room and the side wall, measuring the receiving power. For each change in transmitter-receiver position, received power of 3GHz and 6GHz band were measured and compared with the simulation forecast results. The changes in received power at each receiving point were analyzed according to the location of the transmitter and the frequency band variation.

• Journal of KIISE:Databases
• /
• v.33 no.5
• /
• pp.495-512
• /
• 2006

Recently, researches on a future position prediction of moving objects have been progressed as the importance of the future position retrieval increases. New index structures are required to efficiently retrieve the consecutive positions of moving objects. Existing index structures significantly degrade the search performance of the moving objects because the search operation makes the unnecessary extension of the node in the index structure. To solve this problem, we propose a space partition based index structure considering the mobility of moving objects. To deal with the overflow of a node, our index structure first merges it and the sibling node. If it is impossible to merge them, our method splits the overflow node in which moving properties of objects are considered. Our index structure is always partitioned into overlap free subregions when a node is split. Our split strategy chooses the split position by considering the parameters such as velocities, the escape time of the objects, and the update time of a node. In the internal node, the split position Is determined from preventing the cascading split of the child node. We perform various experiments to show that our index structure outperforms the existing index structures in terms of retrieval performance. Our experimental results show that our proposed index structure achieves about $17%{\sim}264%$ performance gains on current position retrieval and about $107%{\sim}19l%$ on future position retrieval over the existing methods.

• Journal of Korean Society of Water and Wastewater
• /
• v.30 no.2
• /
• pp.179-185
• /
• 2016

In this study, the RDII predictions were compared using two methodologies, i.e., the RTK-based and regression methods. Long-term (1/1/2011~12/31/2011) monitoring data, which consists of 10-min interval streamflow and the amount of precipitation, were collected at the domestic study area (1.36 km2 located in H county), and used for the construction of the RDII prediction models. The RTK method employs super position of tri-triangles, and each triangle (called, unit hydrograph) is defined by three parameters (i.e., R, T and K) determined/optimized using Genetic Algorithm (GA). In regression method, the MovingAverage (MA) filtering was used for data processing. Accuracies of RDII predictions from these two approaches were evaluated by comparing the root mean square error (RMSE) values from each model, in which the values were calculated to 320.613 (RTK method) and 420.653 (regression method), respectively. As a results, the RTK method was found to be more suitable for RDII prediction during extreme rainfall event, than the regression method.

• Journal of Astronomy and Space Sciences
• /
• v.8 no.1
• /
• pp.99-113
• /
• 1991

We developed a software system called IODS(ISSA Orbit Determination System), which can predict the orbit of arbitary artificial satellite using the numerical method. For evaluating the orbit prediction accuracy of IODS, the orbital data predicted for the meteorological satellite NOAA-11 and the stationary satellite INTELSAT-V are intercompared with those tracked at the Central Bureau of Meterology and the Kum-San Satellite Communication Station. And the Perturbations affecting the orbit of these artificial satellites are quantitatively analyzed. The orbital variation and the eclipse phenomina due to the shadow are analyzed for a hypothetical geostationary satellite called KORSAT-1 which is assumed to be located in longitude $110^{circ}E$.