• Journal of KIISE:Computer Systems and Theory
• /
• v.32 no.10
• /
• pp.520-529
• /
• 2005

A feature-based 3D modeling algorithm is presented in this paper. Since conventional methods use depth-based techniques, they need much time for the image matching to extract depth information. Even feature-based methods have less computation load than that of depth-based ones, the calculation of modeling error about whole pixels within a triangle is needed in feature-based algorithms. It also increase the computation time. Therefore, the proposed algorithm consists of three phases, which are an initial 3D model generation, model evaluation, and model refinement phases, in order to acquire an efficient 3D model. Intensity gradients and incremental Delaunay triangulation are used in the Initial model generation. In this phase, a morphological edge operator is adopted for a fast edge filtering, and the incremental Delaunay triangulation is modified to decrease the computation time by avoiding the calculation errors of whole pixels and selecting a vertex at the near of the centroid within the previous triangle. After the model generation, sparse vertices are matched, then the faces are evaluated with the size, approximation error, and disparity fluctuation of the face in evaluation stage. Thereafter, the faces which have a large error are selectively refined into smaller faces. Experimental results showed that the proposed algorithm could acquire an adaptive model with less modeling errors for both smooth and abrupt areas and could remarkably reduce the model acquisition time.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.49 no.9
• /
• pp.347-352
• /
• 2012

In this paper, an 2D spatial-map construction for workers identification and avoidance of AGV using the detection scheme of the spatial coordinates based on stereo camera is proposed. In the proposed system, face area of a moving person is detected from a left image among the stereo image pairs by using the YCbCr color model and its center coordinates are computed by using the centroid method and then using these data, the stereo camera embedded on the mobile robot can be controlled for tracking the moving target in real-time. Moreover, using the disparity map obtained from the left and right images captured by the tracking-controlled stereo camera system and the perspective transformation between a 3-D scene and an image plane, depth map can be detected. From some experiments on AGV driving with 240 frames of the stereo images, it is analyzed that error ratio between the calculated and measured values of the worker's width is found to be very low value of 2.19% and 1.52% on average.

• Atmosphere
• /
• v.25 no.1
• /
• pp.117-127
• /
• 2015

The presences of clouds significantly influence the accuracy of ozone retrievals from satellite measurements. This study focuses on the influence of clouds on Ozone Monitoring instrument (OMI) ozone profile retrieval based on an optimal estimation. There are two operational OMI cloud products; OMCLDO2, based on absorption in $O_2-O_2$ at 477 nm, and OMCLDRR, based on filling in Fraunhofer lines by rotational Raman scattering (RRS) at 350 nm. Firstly, we characterize differences between $O_2-O_2$ and RRS effective cloud pressures using MODIS cloud optical thickness (COT), and then compare ozone profile retrievals with different cloud input data. $O_2-O_2$ cloud pressures are significantly smaller than RRS by ~200 hPa in thin clouds, which corresponds to either low COT or cloud fraction (CF). On the other hand, the effect of Optical centroid pressure (OCP) on ozone retrievals becomes significant at high CF. Tropospheric ozone retrievals could differ by up to ${\pm}10$ DU with the different cloud inputs. The layer column ozone below 300 hPa shows the cloud-induced ozone retrieval error of more than 20%. Finally, OMI total ozone is validated with respect to Brewer ground-based total ozone. A better agreement is observed when $O_2-O_2$ cloud data are used in OMI ozone profile retrieval algorithm. This is distinctly observed at low OCP and high CF.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.5
• /
• pp.185-192
• /
• 2004

This paper presents face recognition using vector quantization in the eigenspace of the faces. The existing eigenface method is not enough for representing the variations of faces. For making up for its defects, the proposed method use a clustering of feature vectors by vector quantization in eigenspace of the faces. In the trainning stage, the face images are transformed the points in the eigenspace by eigeface(eigenvetor) and we represent a set of points for each people as the centroids of vector quantizer. In the recognition stage, the vector quantizer finds the centroid having the minimum quantization error between feature vector of input image and centriods of database. The experiments are performed by 600 faces in Faces94 database. The existing eigenface method has minimum 64 miss-recognition and the proposed method has minimum 20 miss-recognition when we use 4 codevectors. In conclusion, the proposed method is a effective method that improves recognition rate through overcoming the variation of faces.

• The KIPS Transactions:PartB
• /
• v.10B no.6
• /
• pp.587-592
• /
• 2003

Many real world control problems have continuous states and actions. When the state space is continuous, the reinforcement learning problems involve very large state space and suffer from memory and time for learning all individual state-action values. These problems need function approximators that reason action about new state from previously experienced states. We introduce Fuzzy Q-Map that is a function approximators for 1 - step Q-learning and is based on fuzzy clustering. Fuzzy Q-Map groups similar states and chooses an action and refers Q value according to membership degree. The centroid and Q value of winner cluster is updated using membership degree and TD(Temporal Difference) error. We applied Fuzzy Q-Map to the mountain car problem and acquired accelerated learning speed.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.37 no.3_4
• /
• pp.34-47
• /
• 1995

It is experienced fact as a regular annual event that the structure to he designed on unreasonable flood for the agricultural structures including reservoirs have been brought not only loss of lives, but also enormous property damage. For the solution of this problem at issue, this study was conducted to develop an optimal runoff hydrograph model by comparison of the peak flows and time to peak between observed and simulated flows derived by linear time-invariant and linear time-variant models under the condition of having a short duration of heavy rainfall with uniform rainfall intensity at nine small watersheds which are within the range of 55.9 to 140.7 square kilometers in area in Han, Geum, Nagdong and Yeongsan Rivers. The results obtained through this study can be summarized as follows. 1. Storage constants and Gamma function arguments were calculated within the range of 1.2 to 6.42 and of 1.28 to 8.05 respectively by the moment method as the parameters for the analysis of runoff hydrograph based on linear time-invariant model. 2. Parameters for both linear time-invariant and linear time-variant models were calibrated with nine gaged watershed data, using a trial and error method. The resulting parameters including Gamma function argument, N and storage constant, K for linear time-invariant model were related statistically to watershed characteristic variables such as area, slope, length of main stream and the centroid length of the basin. 3. Average relative errors of the simulated peak discharge of calibrated runoff hydrographs by using linear time-variant and linear time-invariant models were shown to be 0.75 and 5.42 percent respectively to the peak of observed runoff hydrographs. Correlation coefficients for the statistical analysis in the same condition were shown to be 0.999 and 0.978 with a high significance respectively. Therefore, it can be concluded that the accuracy of a linear time-variant model is approaching more closely to the observed runoff hydrograph than that of a linear time-invariant model in the applied watersheds. 4. Average relative errors of the time to peak of calibrated runoff hydrographs by using linear time-variant and linear time-invariant models were shown to be 16.44 and 19.89 percent respectively to the time to peak of observed runoff hydrographs. Correlation coefficients in the same condition were also shown to be 0.999 and 0.886 with a high significance respectively. 5. It can be seen that the shape of simulated hydrograph based on a linear time- variant model is getting closer to the observed runoff hydrograph than that of a linear time-invariant model in the applied watersheds. 6. Two different models were verified with different rainfall-runoff events from data for the calibration by relative error and correlation analysis. Consequently, it can be generally concluded that verification results for the peak discharge and time to peak of simulated runoff hydrographs were in good agreement with those of calibrated runoff hydrographs.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.1C
• /
• pp.26-35
• /
• 2006

In this paper, an automatic mobile robot system for a intelligent path planning using the detection scheme of the spatial coordinates based on stereo camera is proposed. In the proposed system, face area of a moving person is detected from a left image among the stereo image pairs by using the YCbCr color model and its center coordinates are computed by using the centroid method and then using these data, the stereo camera embedded on the mobile robot can be controlled for tracking the moving target in real-time. Moreover, using the disparity map obtained from the left and right images captured by the tracking-controlled stereo camera system and the perspective transformation between a 3-D scene and an image plane, depth information can be detected. Finally, based-on the analysis of these calculated coordinates, a mobile robot system is derived as a intelligent path planning and a estimation. From some experiments on robot driving with 240 frames of the stereo images, it is analyzed that error ratio between the calculated and measured values of the distance between the mobile robot and the objects, and relative distance between the other objects is found to be very low value of $2.19\%$ and $1.52\%$ on average, respectably.

• The KIPS Transactions:PartB
• /
• v.14B no.7
• /
• pp.493-502
• /
• 2007

An on-line registration technique is presented to register multi-view range images for the 3D reconstruction of real objects. Using a range camera, we first acquire range images and photometric images continuously. In the range images, we divide object and background regions using a predefined threshold value. For the coarse registration of the range images, the centroid of the images are used. After refining the registration of range images using a projection-based technique, we use a modified KLT(Kanade-Lucas-Tomasi) tracker to match photometric features in the object images. Using the modified KLT tracker, we can track image features fast and accurately. If a range image fails to register, we acquire new range images and try to register them continuously until the registration process resumes. After enough range images are registered, they are integrated into a 3D model in offline step. Experimental results and error analysis show that the proposed method can be used to reconstruct 3D model very fast and accurately.

• Nuclear Medicine and Molecular Imaging
• /
• v.42 no.1
• /
• pp.61-69
• /
• 2008

Purpose: Micro-pinhole SPECT system with conventional multiple-head gamma cameras has the advantage of high magnification factor for imaging of rodents. However, several geometric factors should be calibrated to obtain the SPECT image with good image quality. We developed a simplified geometric calibration method for rotating triple-head pinhole SPECT system and assessed the effects of the calibration using several phantom and rodent imaging studies. Materials and Methods: Trionix Triad XLT9 triple-head SPECT scanner with 1.0 mm pinhole apertures were used for the experiments. Approximately centered point source was scanned to track the angle-dependent positioning errors. The centroid of point source was determined by the center of mass calculation. Axially departed two point sources were scanned to calibrate radius of rotation from pinhole to center of rotation. To verify the improvements by the geometric calibration, we compared the spatial resolution of the reconstructed image of Tc-99m point source with and without the calibration. SPECT image of micro performance phantom with hot rod inserts was acquired and several animal imaging studies were performed. Results: Exact sphere shape of the point source was obtained by applying the calibration and axial resolution was improved. Lesion detectibility and image quality was also much improved by the calibration in the phantom and animal studies. Conclusion: Serious degradation of micro-pinhole SPECT images due to the geometric errors could be corrected using a simplified calibration method using only one or two point sources.