• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.9
• /
• pp.2605-2625
• /
• 2024

Siamese-based segmentation and tracking algorithms improve accuracy and stability for video object segmentation and tracking tasks simultaneously. Although effective, variability in target appearance and background clutter can still affect segmentation accuracy and further influence the performance of tracking. In this paper, we present a memory propagation-based target-aware and mask-attention decision network for robust object segmentation and tracking. Firstly, a mask propagation-based attention module (MPAM) is constructed to explore the inherent correlation among image frames, which can mine mask information of the historical frames. By retrieving a memory bank (MB) that stores features and binary masks of historical frames, target attention maps are generated to highlight the target region on backbone features, thus suppressing the adverse effects of background clutter. Secondly, an attention refinement pathway (ARP) is designed to further refine the segmentation profile in the process of mask generation. A lightweight attention mechanism is introduced to calculate the weight of low-level features, paying more attention to low-level features sensitive to edge detail so as to obtain segmentation results. Finally, a mask fusion mechanism (MFM) is proposed to enhance the accuracy of the mask. By utilizing a mask quality assessment decision network, the corresponding quality scores of the "initial mask" and the "previous mask" can be obtained adaptively, thus achieving the assignment of weights and the fusion of masks. Therefore, the final mask enjoys higher accuracy and stability. Experimental results on multiple benchmarks demonstrate that our algorithm performs outstanding performance in a variety of challenging tracking tasks.

• Journal of Korean Society of Occupational and Environmental Hygiene
• /
• v.32 no.2
• /
• pp.111-115
• /
• 2022

Objectives: The purpose of this study is to verify whether droplet-induced propagation, the main route of infectious diseases such as COVID-19, can occur in semiconductor FAB (Fabrication), based on research results on general droplet propagation. Methods: Through data surveys droplet propagation was modeled through simulation and experimental case analysis according to general (without mask) and mask-wearing conditions, and the risk of droplet propagation was inferred by reflecting semiconductor FAB operation conditions (air current, air conditioning system, humidity, filter conditions). Results: Based on the results investigated to predict the possibility of spreading infectious diseases in semiconductor FAB, the total amount of droplet propagation (concentration), propagation distance, and virus life in FAB were inferred by reflecting the management parameter of semiconductor FAB. Conclusions: The total amount(concentration) of droplet propagation in the semiconductor fab is most affected by the presence or absence of wearing a mask and the line air dilution rate has some influence. when worn it spreads within 0.35~1m, and since the humidity is constant the virus can survive in the air for up to 3 hours. as a result the semiconductor fab is judged to be and effective space to block virus propagation due to the special environmental condition of a clean room.

• Current Optics and Photonics
• /
• v.7 no.1
• /
• pp.1-14
• /
• 2023

Wavefront-coding imaging can achieve high-quality imaging along with a wide range of defocus. In this paper, the anti-laser detection and damage performance of wavefront-coding imaging systems using different asymmetric phase masks are studied, through modeling and simulation. Based on FresnelKirchhoff diffraction theory, the laser-propagation model of the wavefront-coding imaging system is established. The model uses defocus distance rather than wave aberration to characterize the degree of defocus of an imaging system. Then, based on a given defocus range, an optimization method based on Fisher information is used to determine the optimal phase-mask parameters. Finally, the anti-laser detection and damage performance of asymmetric phase masks at different defocus distances and propagation distances are simulated and analyzed. When studying the influence of defocus distance, compared to conventional imaging, the maximum single-pixel receiving power and echo-detection receiving power of asymmetric phase masks are reduced by about one and two orders of magnitude respectively. When exploring the influence of propagation distance, the maximum single-pixel receiving power of asymmetric phase masks decreases by about one order of magnitude and remains stable, and the echodetection receiving power gradually decreases with increasing propagation distance, until it approaches zero.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.18 no.4
• /
• pp.34-42
• /
• 1981

A 512-bit Task Programmable ROM has been designed and fabricated using PMOS technology. The content of the memory was written through the gate pattern during the fabrication process, and was checked by displaying the output of the chip on an oscilloscope with 512(32$\times$16) matrix points. The operation of the chip was surcessful with operating voltage from -6V to -l2V, The power consumption and propagation delay time have been measured to be 3mW and 13 $\mu$sec, respectively at -6 Volt. The power consunption increased to 27mW and propagation delay time decreased to 3$\mu$sec at -12V. The output of the chip was capable of driving the input of a TTL gate directly and retained a high impedence state when the chip solect function disabled the output.

• Journal of Institute of Control, Robotics and Systems
• /
• v.9 no.5
• /
• pp.361-367
• /
• 2003

We propose an algorithm to detect the face region using a variable ellipsoidal mask and a neural network. Since outlines of human faces are similar to ellipsoid, the ellipsoidal mask that has the fixed ratio of major and minor axis can be used to detect the candidate area. The positions of eyes and lips are extracted in this candidate area, and then the morphological analysis is applied to make features which are consist of six parameters, such as the geometrical ratio of eyes and lips. A back-propagation neural network is used as a classifier to determine the most possible face region. The experimental result is conducted to verify its efficiency compared with those of previous works.

• Proceedings of the IEEK Conference
• /
• 2002.06c
• /
• pp.87-90
• /
• 2002

This paper has been studied a new approach using neural network to detect false laser marks. In the proposed approach, input images are segmented into R, G and B colors and implements mask areas respectively. And then average and variation values of the each mask area are extracted for the learning process to minimize input nodes. Using this technique, the new input data is obtained and implemented to the back-propagation algorithm using multi layer perception. This paper reduces the computational complexity necessary and shows better effectiveness to inspect false laser marks.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.47 no.1
• /
• pp.9-16
• /
• 2010

Moving object detection is to detect foreground object different from background scene in a new incoming image frame and is an essential ingredient process in some image processing applications such as intelligent visual surveillance, HCI, object-based video compression and etc. Most of previous object detection algorithms are still computationally heavy so that it is difficult to develop real-time multi-channel moving object detection in a workstation or even one-channel real-time moving object detection in an embedded system using them. Foreground mask correction necessary for a more precise object detection is usually accomplished using morphological operations like opening and closing. Morphological operations are not computationally cheap and moreover, they are difficult to be rendered to run simultaneously with the subsequent connected component labeling routine since they need quite different type of processing from what the connected component labeling does. In this paper, we first devise a fast and precise foreground mask correction algorithm, "Neighbor Foreground Pixel Propagation (NFPP)" which utilizes neighbor pixel checking employed in the connected component labeling. Next, we propose a novel moving object detection method based on the devised foreground mask correction algorithm, NFPP where the connected component labeling routine can be executed simultaneously with the foreground mask correction. Through experiments, it is verified that the proposed moving object detection method shows more precise object detection and more than 4 times faster processing speed for a image frame and videos in the given the experiments than the previous moving object detection method using morphological operations.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.10 no.2
• /
• pp.134-142
• /
• 2010

A comparative study on the trade-off between the drive current and the total gate capacitance in double-gate (DG) and triple-gate (TG) FinFETs is performed by using 3-D device simulation. As the first result, we found that the optimum ratio of the hardmask oxide thickness ($T_{mask}$) to the sidewall oxide thickness ($T_{ox}$) is $T_{mask}/T_{ox}$=10/2 nm for the minimum logic delay ($\tau$) while $T_{mask}/T_{ox}$=5/1~2 nm for the maximum intrinsic gate capacitance coupling ratio (ICR) with the fixed channel length ($L_G$) and the fin width ($W_{fin}$) under the short channel effect criterion. It means that the TG FinFET is not under the optimal condition in terms of the circuit performance. Second, under optimized $T_{mask}/T_{ox}$, the propagation delay ($\tau$) decreases with the increasing fin height $H_{fin}$. It means that the FinFET-based logic circuit operation goes into the drive current-dominant regime rather than the input gate load capacitance-dominant regime as $H_{fin}$ increases. In the end, the sensitivity of $\Delta\tau/{\Delta}H_{fin}$ or ${{\Delta}I_{ON}}'/{\Delta}H_{fin}$ decreases as $L_G/W_{fin}$ is scaled-down. However, $W_{fin}$ should be carefully designed especially in circuits that are strongly influenced by the self-capacitance or a physical layout because the scaling of $W_{fin}$ is followed by the increase of the self-capacitance portion in the total load capacitance.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2003.06a
• /
• pp.434-437
• /
• 2003

Evaluation or the patterning accuracy of e-beam lithography machines requires a high precision inspection system that is capable of measuring the true xy-locations of fiducial marks generated by the e-beam machine under test. Fiducial marks are fabricated on a single photo mask over the entire working area in the form of equally spaced two-dimensional grids. In performing the evaluation, the principles of self-calibration enable to determine the deviations of fiducial marks from their nominal xy-locations precisely, not being affected by the motion errors of the inspection system itself. It is. however, the fact that only repeatable motion errors can be eliminated, while random motion errors encountered in probing the locations of fiducial marks are not removed. Even worse, a random error occurring from the measurement of a single mark propagates and affects in determining locations of other marks, which phenomenon in fact limits the ultimate calibration accuracy of e-beam machines. In this paper, we describe an uncertainty analysis that has been made to investigate how random errors affect the final result of self-calibration of e-beam machines when one uses an optical inspection system equipped with high-resolution microscope objectives and a precision xy-stages. The guide of uncertainty analysis recommended by the International Organization for Standardization is faithfully followed along with necessary sensitivity analysis. The uncertainty analysis reveals that among the dominant components of the patterning accuracy of e-beam lithography, the rotationally symmetrical component is most significantly affected by random errors, whose propagation becomes more severe in a cascading manner as the number of fiducial marks increases

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.47 no.4
• /
• pp.20-25
• /
• 2010

An integrated Mach-Zehnder interferometric sensor operating at 632.8 nm was designed and fabricated by the technology of planar rib waveguides. Rib waveguide based on silica system ($SiO_2-SiO_xN_y-SiO_2$) was geometrically designed to have single mode operation and high sensitivity. It was structured by semiconductor fabrication processes such as thin film deposition, photolithography, and RIE (Reactive Ion Etching). With the power observation, propagation loss measurement by cut-back method showed about 4.82 dB/cm for rib waveguides. Additionally the chromium mask process for an etch stop was employed to solve the core damaging problem in patterning the sensing zone on the chip. Refractive index measurement of water/ethanol mixture with this device finally showed a sensitivity of about $\pi$/($4.04{\times}10^{-3}$).