• Journal of Advanced Marine Engineering and Technology
• /
• v.35 no.7
• /
• pp.930-937
• /
• 2011

The characteristics of mean velocity and turbulence have been analyzed in the circular open channel flow using PIV measurement data for a wide range of water depth. The measured data are fitted to a velocity distribution function over the whole depth of the open channel. Reynolds shear stress and mean velocity in wall unit are compared with the analytic models for fully-developed turbulent boundary layer. Both the mean velocity and Reynolds shear stress have different distributions from the two-dimensional boundary layer flow when the water depth increases over 50% since the influence of the side wall penetrates more deeply into the free surface. The cross-stream Reynolds normal stress also has considerably different distribution in view of its peak value and decreasing rate in the outer region whether the water depth is higher than 50% or not.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.1
• /
• pp.67-74
• /
• 2023

The images acquired from the speaker's mouth region revealed the unique patterns according to the corresponding voices. By using this principle, the several methods were proposed in which speech signals were recognized or synthesized from the images acquired at the speaker's lower face. In this study, an image-based speech synthesis method was proposed in which the depth images were cooperatively used. Since depth images yielded depth information that cannot be acquired from optical image, it can be used for the purpose of supplementing flat optical images. In this paper, the usefulness of depth images from the perspective of speech synthesis was evaluated. The validation experiment was carried out on 60 Korean isolated words, it was confirmed that the performance in terms of both subjective and objective evaluation was comparable to the optical image-based method. When the two images were used in combination, performance improvements were observed compared with when each image was used alone.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.7
• /
• pp.1706-1725
• /
• 2024

The unique U-shaped structure of U-Net network makes it achieve good performance in image segmentation. This network is a lightweight network with a small number of parameters for small image segmentation datasets. However, when the medical image to be segmented contains a lot of detailed information, the segmentation results cannot fully meet the actual requirements. In order to achieve higher accuracy of medical image segmentation, a novel improved U-Net network architecture called multi-scale encoder-decoder U-Net+ (MEDU-Net+) is proposed in this paper. We design the GoogLeNet for achieving more information at the encoder of the proposed MEDU-Net+, and present the multi-scale feature extraction for fusing semantic information of different scales in the encoder and decoder. Meanwhile, we also introduce the layer-by-layer skip connection to connect the information of each layer, so that there is no need to encode the last layer and return the information. The proposed MEDU-Net+ divides the unknown depth network into each part of deconvolution layer to replace the direct connection of the encoder and decoder in U-Net. In addition, a new combined loss function is proposed to extract more edge information by combining the advantages of the generalized dice and the focal loss functions. Finally, we validate our proposed MEDU-Net+ MEDU-Net+ and other classic medical image segmentation networks on three medical image datasets. The experimental results show that our proposed MEDU-Net+ has prominent superior performance compared with other medical image segmentation networks.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.12 no.1
• /
• pp.27-33
• /
• 2016

In this study we propose a method for 3D landform reconstruction and object modeling method by generating contour lines on the map using a depth sensor which abstracts characteristics of geological layers from the depth map. Unlike the common visual camera, the depth-sensor is not affected by the intensity of illumination, and therefore a more robust contour and object can be extracted. The algorithm suggested in this paper first abstracts the characteristics of each geological layer from the depth map image and rearranges it into the proper order, then creates contour lines using the Bezier curve. Using the created contour lines, 3D images are reconstructed through rendering by mapping RGB images of the visual camera. Experimental results show that the proposed method using depth sensor can reconstruct contour map and 3D modeling in real-time. The generation of the contours with depth data is more efficient and economical in terms of the quality and accuracy.

• Current Optics and Photonics
• /
• v.5 no.1
• /
• pp.23-31
• /
• 2021

We propose a wheel screen type Lamina 3D display, which realizes a 3D image that can satisfy the accommodation cue by projecting volumetric images encoded by varying polarization states to a multilayered screen. The proposed system is composed of two parts: an encoding part that converts depth information to states of polarization and a decoding part that projects depth images to the corresponded diffusing layer. Though the basic principle of Lamina displays has already been verified by previous studies, those schemes suffered from a bottleneck of inferior resolution of the 3D image due to the blurring on the surfaces of diffusing layers in the stacked volume. In this paper, we propose a new structure to implement the decoding part by adopting a form of the wheel screen. Experimental verification is also provided to support the proposed principle.

• Journal of Korea Multimedia Society
• /
• v.20 no.2
• /
• pp.225-233
• /
• 2017

Line-laser beams are used for accurate measurement of 3D shape, which is robust to external illumination. For depth measurement, we project a line-laser beam across an object from the face and take an image of the beam on the object surface using a CCD camera at some angle with respect to the face. For shape measurement, we project parallel line-laser beams with narrow line to line distance. When a layer of thin materials attached to a cylinder is long narrow along its circumference, we can measure the shape of the layer with a small number of parallel line beams if we project line beams along the circumference of the cylinder. Measurement of the depth of the attached materials on a line-laser beam is based on the number of pixels between an imaginary line along the imaginary cylinder without the attached materials and the beam line along the materials attached to the cylinder. For this we need to localize the imaginary line in the captured image. In this paper, we model the shape of the line as an ellipse and localize the line with least square estimate. The proposed method results in smaller error (maximum 0.24mm) than a popular 3D depth camera (maximum 1mm).

• Structural Engineering and Mechanics
• /
• v.77 no.1
• /
• pp.47-56
• /
• 2021

This paper deals with damage detection using a Gapped Smoothing Method (GSM) combined with deep learning. Convolutional Neural Network (CNN) is a model of deep learning. CNN has an input layer, an output layer, and a number of hidden layers that consist of convolutional layers. The input layer is a tensor with shape (number of images) × (image width) × (image height) × (image depth). An activation function is applied each time to this tensor passing through a hidden layer and the last layer is the fully connected layer. After the fully connected layer, the output layer, which is the final layer, is predicted by CNN. In this paper, a complete machine learning system is introduced. The training data was taken from a Finite Element (FE) model. The input images are the contour plots of curvature gapped smooth damage index. A free-free beam is used as a case study. In the first step, the FE model of the beam was used to generate data. The collected data were then divided into two parts, i.e. 70% for training and 30% for validation. In the second step, the proposed CNN was trained using training data and then validated using available data. Furthermore, a vibration experiment on steel damaged beam in free-free support condition was carried out in the laboratory to test the method. A total number of 15 accelerometers were set up to measure the mode shapes and calculate the curvature gapped smooth of the damaged beam. Two scenarios were introduced with different severities of the damage. The results showed that the trained CNN was successful in detecting the location as well as the severity of the damage in the experimental damaged beam.

• Journal of Broadcast Engineering
• /
• v.15 no.1
• /
• pp.3-13
• /
• 2010

In this paper, we propose a method for filling out the occlusions in generating multi-view images from one source image and its ground-truth depth image. The method is based on image inpainting and layered interpolation. The source image is first divided into several layers using depth information. The occlusions are interpolated separately in every layered image using the image inpainting algorithm. Finally, the interpolated layered images are combined to obtain different viewpoint images. Interpolating occlusions with depth-correlated texture information that is contained to each layer makes it possible to obtain more detailed and accurate results than previous methods. The effectiveness of the proposed method is shown through experimental results.

• Journal of the Korean Society of Radiology
• /
• v.13 no.3
• /
• pp.319-324
• /
• 2019

A depth-encoding detector module with silicon photomultipliers(SiPMs) using two layers of scintillation crystal array was designed, and the position measurement capability was verified using DETECT2000. The depth of interaction of the crystal pixels with the gamma rays was tracked through the image acquired with the combination of surface treatment of the crystal pixels and reflectors. The bottom layer was treated as a reflector except for the optically coupled surfaces, and the crystals of top layer were optically coupled each other except for the outer surfaces so that the light sharing was made easier than the bottom layer. Flood images were obtained through the combination of specular reflectors and random reflectors, grounded and polished surfaces of crystal pixels, and the positions at which layer images were generated were measured and analyzed. The images were reconstructed using the Anger algorithm, whose the SiPM signals were reduced as the 16-channels to 4-channels. In the combination of the grounded surface and all reflectors, the depth positions were discriminated into two layers, whereas it was impossible to separate the two layers in the all polished surface combinations. Therefore, using the combination of grounded surface crystal pixels and reflectors could improve the spatial resolution at the outside of the field of view by measuring the depth position in preclinical positron emission tomography.

• The Journal of Korea Robotics Society
• /
• v.17 no.4
• /
• pp.417-424
• /
• 2022

In this paper, the developed trunk cargo unloading automation system is introduced, and the RGB-D sensor-based box loading situation recognition method and unloading plan applied to this system are suggested. First of all, it is necessary to recognize the position of the box in a truck. To do this, we first apply CNN-based YOLO, which can recognize objects in RGB images in real-time. Then, the normal vector of the center of the box is obtained using the depth image to reduce misrecognition in parts other than the box, and the inner wall of the truck in an image is removed. And a method of classifying the layers of the boxes according to the distance using the recognized depth information of the boxes is suggested. Given the coordinates of the boxes on the nearest layer, a method of generating the optimal path to take out the boxes the fastest using this information is introduced. In addition, kinematic analysis is performed to move the conveyor to the position of the box to be taken out of the truck, and kinematic analysis is also performed to control the robot arm that takes out the boxes. Finally, the effectiveness of the developed system and algorithm through a test bed is proved.