• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.6
• /
• pp.28-34
• /
• 2020

This paper proposes an approach to the fusion of two heterogeneous sensors with two different fields-of-view (FOV): LIDAR and an RGB camera. Registration between data captured by LIDAR and an RGB camera provided the fusion results. Registration was completed once a depthmap corresponding to a 2-dimensional RGB image was generated. For this fusion, RPLIDAR-A3 (manufactured by Slamtec) and a general digital camera were used to acquire depth and image data, respectively. LIDAR sensor provided distance information between the sensor and objects in a scene nearby the sensor, and an RGB camera provided a 2-dimensional image with color information. Fusion of 2D image and depth information enabled us to achieve better performance with applications of object detection and tracking. For instance, automatic driver assistance systems, robotics or other systems that require visual information processing might find the work in this paper useful. Since the LIDAR only provides depth value, processing and generation of a depthmap that corresponds to an RGB image is recommended. To validate the proposed approach, experimental results are provided.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.6
• /
• pp.11-19
• /
• 2019

In this paper, a new Scene-based Nonuniformity Correction (SBNUC) method is proposed by applying Image Roughness-like and Spatial Noise cost functions on deep neural network structure. The classic approaches for nonuniformity correction require generally plenty of sequential image data sets to acquire accurate image correction offset coefficients. The proposed method, however, is able to estimate offset from only a couple of images powered by the characteristic of deep neural network scheme. The real world SWIR image set is applied to verify the performance of proposed method and the result shows that image quality improvement of PSNR 70.3dB (maximum) is achieved. This is about 8.0dB more than the improved IRLMS algorithm which preliminarily requires precise image registration process on consecutive image frames.

• Journal of the Institute of Convergence Signal Processing
• /
• v.3 no.4
• /
• pp.29-34
• /
• 2002

In this paper, we present a new geometric active contour model based on level set methods introduced by Osher and Sethian for detection of object boundaries or shape and we adopt anisotropic diffusion filtering method for removing noise from original image. In order to minimize the processing time, we use the narrow band method which allows us to perform calculations in the neighborhood of the contour and not in the whole image. Using anisotropic diffusion filtering for each slice, we have the result with reduced noise and extracted exact shape. Volume rendering operates on three-dimensional data, processes it, and transforms it into a simple two-dimensional image.

• Progress in Medical Physics
• /
• v.27 no.4
• /
• pp.203-212
• /
• 2016

Dose differences between three-dimensional (3D) and four-dimensional (4D) doses could be varied according to the geometrical relationship between a planning target volume (PTV) and an organ at risk (OAR). The purpose of this study is to evaluate the correlation between the overlap volume histogram (OVH), which quantitatively shows the geometrical relationship between the PTV and OAR, and the dose differences. 4D computed tomography (4DCT) images were acquired for 10 liver cancer patients. Internal target volume-based treatment planning was performed. A 3D dose was calculated on a reference phase (end-exhalation). A 4D dose was accumulated using deformation vector fields between the reference and other phase images of 4DCT from deformable image registration, and dose differences between the 3D and 4D doses were calculated. An OVH between the PTV and selected OAR (duodenum) was calculated and quantified on the basis of specific overlap volumes that corresponded to 10%, 20%, 30%, 40%, and 50% of the OAR volume overlapped with the expanded PTV. Statistical analysis was performed to verify the correlation with the OVH and dose difference for the OAR. The minimum mean dose difference was 0.50 Gy from case 3, and the maximum mean dose difference was 4.96 Gy from case 2. The calculated range of the correlation coefficients between the OVH and dose difference was from -0.720 to -0.712, and the R-square range for regression analysis was from 0.506 to 0.518 (p-value <0.05). However, when the 10% overlap volume was applied in the six cases that had OVH value ${\leq}2$, the average percent mean dose differences were $34.80{\pm}12.42%$. Cases with quantified OVH values of 2 or more had mean dose differences of $29.16{\pm}11.36%$. In conclusion, no significant statistical correlation was found between the OVH and dose differences. However, it was confirmed that a higher difference between the 3D and 4D doses could occur in cases that have smaller OVH value.

• Journal of the Korean Society of Radiology
• /
• v.11 no.4
• /
• pp.235-240
• /
• 2017

The purpose of this study is to confirm the safety of the clinical application of image co - registration in steteotactic radiosurgery by evaluating the 3D positioning of magnetic resonance imaging using image co-registration. We performed a retrospective study using three-dimensional coordinate measurement of 32 patients who underwent stereotactic radiosurgery and performed magnetic resonance imaging follow-up using image co-registration. The 3 dimensional coordinate errors were $1.0443{\pm}0.5724mm$ (0.10 ~ 1.89) in anterior commissure and $1.0348{\pm}0.5473mm$ (0.36 ~ 2.24) in posterior commissure. The mean error of MR1 (3.0 T) was lower than that of MR2 (1.5 T). It is necessary to minimize the error of magnetic resonance imaging in the treatment planning using the image co - registration technique and to confirm it.

• The Academic Congress of Korean Shoulder and Elbow Society
• /
• 2009.03a
• /
• pp.44-44
• /
• 2009

The purpose of this study was to investigate humeral translation relative to the glenoid invivo during loaded and unloaded shoulder abduction. CT scans of 9 healthy shoulders were acquired and 3D models were created. The subject was positioned in front of a fluoroscope and motions were recorded during active abduction. The subjects performed two trials of holding a 3kg weight and unload. 3D motions were determined using model-based 3D-to-2D registration to obtain 6 degrees of freedom kinematics. Glenohumeral translation was determined by finding the location on the humeral head with the smallest separation from the glenoid. Humeral translation was referenced to the glenoid center in the superior/inferior direction. The humerus moved an average of 2 mm, from inferior to central on the glenoid, during arm abduction for both conditions. The humeral head was centered within 1mm from the glenoid center above $70^{\circ}$. There were no statistically significant differences for both conditions. The standard deviation decreased gradually over the motion, with significantly lower variability at the end of abduction compared to the initial unloaded position. We assumed that the humeral translation to the center of the glenoid provides maximum joint congruency for optimal shoulder function and joint longevity. We believe this information will lead to better strategies to prevent shoulder injuries, enhance rehabilitation, and improve surgical treatments.

• Imaging Science in Dentistry
• /
• v.49 no.1
• /
• pp.35-43
• /
• 2019

Purpose: The goal of this study was to assess the accuracy and reliability of a low-cost portable scanner (Scanify) for imaging facial casts compared to a previously validated portable digital stereophotogrammetry device (Vectra H1). This in vitro study was performed using 2 facial casts obtained by recording impressions of the authors, at King's College London Academic Centre of Reconstructive Science. Materials and Methods: The casts were marked with anthropometric landmarks, then digitised using Scanify and Vectra H1. Computed tomography (CT) scans of the same casts were performed to verify the validation of Vectra H1. The 3-dimensional (3D) images acquired with each device were compared using linear measurements and 3D surface analysis software. Results: Overall, 91% of the linear Scanify measurements were within 1 mm of the corresponding reference values. The mean overall surface difference between the Scanify and Vectra images was <0.3mm. Significant differences were detected in depth measurements. Merging multiple Scanify images produced significantly greater registration error. Conclusion: Scanify is a very low-cost device that could have clinical applications for facial imaging if imaging errors could be corrected by a future software update or hardware revision.

• Spatial Information Research
• /
• v.15 no.2
• /
• pp.123-134
• /
• 2007

The building boundary should be registered in the cadastral map for the protection of property using the expansion of the registration items in the cadastral record and the construction of 3D-cadastral information system. In this study, the efficiency of registering the building boundary was suggested by non-target TS survey through the comparison between target TS and non-target TS survey about an accuracy and efficiency. At the result of this research, the RMSE of target TS and non-target TS survey is the X; ${\pm}0.056m$ and Y; ${\pm}0.043m$. Therefore, non-target TS survey shows the high accuracy. Also, the non-target TS survey is more efficient cost on time and personnel than target TS.

• The Journal of Korea Robotics Society
• /
• v.17 no.3
• /
• pp.373-380
• /
• 2022

This paper considers a method of fast correspondence matching for iterative closest point (ICP) algorithm. In robotics, the ICP algorithm and its variants have been widely used for pose estimation by finding the translation and rotation that best align two point clouds. In computational perspectives, the main difficulty is to find the correspondence point on the reference point cloud to each observed point. Jump-table-based correspondence matching is one of the methods for reducing computation time. This paper proposes a method that corrects errors in an existing jump-table-based correspondence matching algorithm. The criterion activating the use of jump-table is modified so that the correspondence matching can be applied to the situations, such as point-cloud registration problems with highly curved surfaces, for which the existing correspondence-matching method is non-applicable. For demonstration, both hardware and simulation experiments are performed. In a hardware experiment using Hokuyo-10LX LiDAR sensor, our new algorithm shows 100% correspondence matching accuracy and 88% decrease in computation time. Using the F1TENTH simulator, the proposed algorithm is tested for an autonomous driving scenario with 2D range-bearing point cloud data and also shows 100% correspondence matching accuracy.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.6
• /
• pp.43-57
• /
• 2009

Integrating 3D data acquired in multiple views is one of the most important techniques in 3D modeling. However, the existing integration methods are sensitive to registration errors and surface scanning noise. In this paper, we propose a integration algorithm using the local surface topology. We first find all boundary vertex pairs satisfying a prescribed geometric condition in the areas between neighboring surfaces, and then separates areas to several regions by using boundary vertex pairs. We next compute best fitting planes suitable to each regions through PCA(Principal Component Analysis). They are used to produce triangles that be inserted into empty areas between neighboring surfaces. Since each regions between neighboring surfaces can be integrated by using local surface topology, a proposed method is robust to registration errors and surface scanning noise. We also propose a method integrating of textures by using parameterization technique. We first transforms integrated surface into initial viewpoints of each surfaces. We then project each textures to transformed integrated surface. They will be then assigned into parameter domain for integrated surface and be integrated according to the seaming lines for surfaces. Experimental results show that the proposed method is efficient to face modeling.