• Journal of Aerospace System Engineering
• /
• v.16 no.4
• /
• pp.109-115
• /
• 2022

The CNN (Convolution Neural Network) algorithm which combines a deep learning technique, and a computer vision technology, makes image classification feasible with the high-performance computing system. In this thesis, the CNN algorithm is applied to the classification problem, by using a typical deep learning framework of TensorFlow and machine learning techniques. The data set required for supervised learning is generated with the same type of bolts. some of which have undamaged threads, but others have damaged threads. The learning model with less quantity data showed good classification performance on detecting damage in a bolt image. Additionally, the model performance is reviewed by altering the quantity of convolution layers, or applying selectively the over and under fitting alleviation algorithm.

• Journal of Broadcast Engineering
• /
• v.27 no.4
• /
• pp.487-498
• /
• 2022

Depth map is an image that contains distance information in 3D space on a 2D plane and is used in various 3D vision tasks. Many existing depth estimation studies mainly use narrow FoV images, in which a significant portion of the entire scene is lost. In this paper, we propose a technique for generating 360° omnidirectional RGBD images from a sparse set of narrow FoV images. The proposed generative adversarial network based image generation model estimates the relative FoV for the entire panoramic image from a small number of non-overlapping images and produces a 360° RGB and depth image simultaneously. In addition, it shows improved performance by configuring a network reflecting the spherical characteristics of the 360° image.

• Smart Structures and Systems
• /
• v.31 no.4
• /
• pp.351-363
• /
• 2023

Visual structural inspections are an inseparable part of post-earthquake damage assessments. With unmanned aerial vehicles (UAVs) establishing a new frontier in visual inspections, there are major computational challenges in processing the collected massive amounts of high-resolution visual data. We propose twin deep learning models that can provide accurate high-resolution structural components and damage segmentation masks efficiently. The traditional approach to cope with high memory computational demands is to either uniformly downsample the raw images at the price of losing fine local details or cropping smaller parts of the images leading to a loss of global contextual information. Therefore, our twin models comprising Trainable Resizing for high-resolution Segmentation Network (TRS-Net) and DmgFormer approaches the global and local semantics from different perspectives. TRS-Net is a compound, high-resolution segmentation architecture equipped with learnable downsampler and upsampler modules to minimize information loss for optimal performance and efficiency. DmgFormer utilizes a transformer backbone and a convolutional decoder head with skip connections on a grid of crops aiming for high precision learning without downsizing. An augmented inference technique is used to boost performance further and reduce the possible loss of context due to grid cropping. Comprehensive experiments have been performed on the 3D physics-based graphics models (PBGMs) synthetic environments in the QuakeCity dataset. The proposed framework is evaluated using several metrics on three segmentation tasks: component type, component damage state, and global damage (crack, rebar, spalling). The models were developed as part of the 2^nd International Competition for Structural Health Monitoring.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.12
• /
• pp.509-516
• /
• 2022

Accurate pose prediction of objects in 3D space is an important visual recognition technique widely used in many applications such as scene understanding in both indoor and outdoor environments, robotic object manipulation, autonomous driving, and augmented reality. Most previous works for object pose estimation have the limitation that they require an exact 3D CAD model for each object. Unlike such previous works, this paper proposes a novel neural network model that can predict the poses of unknown objects based on only their RGB color images without the corresponding 3D CAD models. The proposed model can obtain depth maps required for unknown object pose prediction by using an adaptive depth estimator, AdaBins,. In this paper, we evaluate the usefulness and the performance of the proposed model through experiments using benchmark datasets.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.17 no.2
• /
• pp.291-298
• /
• 2022

A generative adversarial network (GAN) is a network in which two internal neural networks (generative network and discriminant network) learn while competing with each other. The generator creates an image close to reality, and the delimiter is programmed to better discriminate the image of the constructor. This technology is being used in various ways to create, transform, and restore the entire image X into another image Y. This paper describes a method that can be forged into another object naturally, after extracting only a partial image from the original image. First, a new image is created through the previously trained DCGAN model, after extracting only a partial image from the original image. The original image goes through a process of naturally combining with, after re-styling it to match the texture and size of the original image using the overall style transfer technique. Through this study, the user can naturally add/transform the desired object image to a specific part of the original image, so it can be used as another field of application for creating fake images.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.6
• /
• pp.1023-1030
• /
• 2023

FMCW LiDAR system with robust target detection capabilities even under adverse operating conditions such as snow, rain, and fog is addressed in this paper. Our focus is primarily on enhancing the performance of FMCW LiDAR by improving the characteristics of the frequency-modulated laser, which directly influence range resolution, coherence length, and maximum measurement range etc. of LiDAR. We describe the utilization of an unbalanced Mach-Zehnder laser interferometer to measure real-time changes of the lasing frequency and to correct frequency modulation errors through an optical phase-locked loop technique. To extend the coherence length of laser, we employ an extended-cavity laser diode as the laser source and implement a laser interferometer with an photonic integrated circuit for miniaturization of optical system. The developed FMCW LiDAR system exhibits a bandwidth of 10.045GHz and a remarkable distance resolution of 0.84mm.

• Analytical Science and Technology
• /
• v.37 no.1
• /
• pp.63-78
• /
• 2024

The continuous increasing of the global demand of copper and nickel metals raises the interest in developing alternative technologies to produce them from copper sulfide ore. Also, in line with Egypt's vision 2030 for achieving the sustainable socioeconomic development which aims at developing alternative and eco-friendly technologies for processing the Egyptian ores to produce these strategic products instead of its importing. These metals enhance the advanced electrical and electronic industries. The current work aims at investigating the recovery of copper and nickel from Abu Swayeil copper ore using pug leaching technique by sulfuric acid. The factors affecting the pug leaching process including the sulfuric acid concentration, leaching time and temperature have been investigated. The copper ore sample was characterized chemically using X-ray fluorescence (XRF) and scanning electron microscope (SEM-EDX). A response surface methodology develops a quadratic model that expects the nickel and copper leaching effectiveness as a function of three controlling factors involved in the procedure of leaching was also investigated. The obtained results showed that the maximum dissolution efficiency of Ni and Cu are 99.06 % and 95.30%, respectively which was obtained at the following conditions: 15 % H₂SO₄ acid concentration for 6 hr. at 250 ℃. The dissolution kinetics of nickel and copper that were examined according to heterogeneous model, indicated that the dissolution rates were controlled by surface chemical process during the pug leaching. The activation energy of copper and nickel dissolution were 26.79 kJ.mol^-1 and 38.078 kJ.mol^-1 respectively; and the surface chemical was proposed as the leaching rate-controlling step.

• Investigative Magnetic Resonance Imaging
• /
• v.1 no.1
• /
• pp.103-107
• /
• 1997

Purpose: To investigate the localization and functional lateralization of the supplementary motor area (SMA) in motor activation tests in comparison to that of the primary motor area. Materials and Methods: Seven healthy volunteers obtained echoplanar imaging blood oxygen level dependent technique. This study was carried on 1.5T Siemens Magnetom Vision system with the standard head coil. Parameters of EPI were followed as; TR/TE : 1.0/66.0msec, flip angle: $90^{\circ}$, field of view: $22cm{\times}22cm,{\;}matrix:{\;}128{\times}128$, slice number/slice thickness/gap: 1O/4mm/0.8mm with fat suppression technique. Motor task as finger opposition in each hand consisted of 3 sets of alternative rest and activation periods. Postprocessing were done on Stimulate 5.0 by using cross-correlation statistics. To compare the functional lateralization of the SMA in the right and left hand tests, each examination was evaluated for the percent change of signal intensity and the number of activated voxels both in the SMA and in the pri¬mary motor area. Hemispheric asymmetry was defined as difference of summation of the activted voxels between each hemisphere. Results: Percent change of signal intensity in the SMA (2.49 -3.06%) is lower than that of primary motor area(4.4 -7.23%). Percent change of signal intensity including activated voxels were observed almost equally in the right and left SMA. As for summation of activated voxels, primary motor area had significant difference between each hemisphere but not did the SMA. Conclusion: Preferred contralateral dominant hemisphere and hemispheric asymmetry were detected in the primary motor area but not in the SMA.

• Journal of Korean Ophthalmic Optics Society
• /
• v.16 no.3
• /
• pp.265-272
• /
• 2011

Purpose: The clinical usefulness of rotation evaluation using objective refraction in toric soft lenses fitting was investigated. Methods: Toric soft lenses were fitted for 32 subjects (64 eyes; mean age of 24.69 ${\pm}$ 1.65 years) with astigmatism and both eyes of each subject were fitted with toric soft lenses. Objective refraction-based lenses rotation was evaluated from refraction and over-refraction data by indirect calculating technique. These calculated data were compared with the measured data from slit lamp with direct measuring technique. Results: Orientation of toric soft lenses around zero position (within ${\pm}$ 5$^{\circ}$ vertical line) was investigated. The orientations to the direction of nose of measured and calculated values were 69.78% and 63.64%, respectively, which showed similar values between two techniques. Agreement frequency between measured and calculated values in the magnitude of lenses rotation 54.69% and 82.82% for 10$^{\circ}$ and 20$^{\circ}$ of vertical line, respectively. The 95% limits of agreement between calculation and measurement were from -10.08$^{\circ}$ to 12.65$^{\circ}$, and mean difference was 1.29$^{\circ}$ within ${\pm}$ 10$^{\circ}$. The result showed there was no significant difference (p = 0.1984) and high correlation (r = 0.56, p = 0.0004) between two techniques. But the 95% limits of agreement was widen in ${\pm}$ 20$^{\circ}$ of vertical line. The magnitude of lens rotation between two methods was 9.66 ${\pm}$ 6.16$^{\circ}$, 16.17 ${\pm}$ 12.38$^{\circ}$ and 10.58 ${\pm}$ 12.02$^{\circ}$ for normal, loose and tight fitted conditions. Conclusions: From the results with smaller difference between two techniques, it was found that higher availability of subjective over-refraction data can be used as a supplementary tool for subjective refraction. An application using objective refraction with direct measuring could be provide high success in prescription on toric soft lenses.

• Journal of the Korean Arthroscopy Society
• /
• v.5 no.2
• /
• pp.120-123
• /
• 2001

Purpose : We used transparent flexible tube which had provided the good visual field of median nerve when it was used in endoscopic release of transverse carpal ligament and evaluated the safety of that technique. Materials and Methods : We evaluated the 12 patients(20cases) who had been diagnosed as carpal tunnel syndrome and performed by endoscopic carpal tunnel release between Mar. 1997 and Mar. 2000. We used two portal technique and released the transverse carpal ligament with direct supervision of median nerve. Results : 14 cases$(70\%)$ were revealed excellent or good results and 6 cases$(30\%)$ were fair. No serious complications were shown such as nerve injury. Conclusion : We could avoid the complications of endoscopic carpal tunnel release using the transparent flexible tube which had provided tire good circumferential vision around the median nerve and it is unnecessary to maintain the wrist Joint hyperextension state during operation. Also that tube was easily obtainable in hospital so we need not to purchase the expensive operation apparatus.