• ETRI Journal
• /
• 제43권3호
• /
• pp.538-548
• /
• 2021

This paper proposes an approach to improve the performance of no-reference video quality assessment for sports videos with dynamic motion scenes using an efficient spatiotemporal model. In the proposed method, we divide the video sequences into video blocks and apply a 3D shearlet transform that can efficiently extract primary spatiotemporal features to capture dynamic natural motion scene statistics from the incoming video blocks. The concatenation of a deep residual bidirectional gated recurrent neural network and logistic regression is used to learn the spatiotemporal correlation more robustly and predict the perceptual quality score. In addition, conditional video block-wise constraints are incorporated into the objective function to improve quality estimation performance for the entire video. The experimental results show that the proposed method extracts spatiotemporal motion information more effectively and predicts the video quality with higher accuracy than the conventional no-reference video quality assessment methods.

• 대한건축학회논문집:구조계
• /
• 제35권3호
• /
• pp.35-41
• /
• 2019

In this study, the fire resistance performance of polymer cement mortars which are used as a representative repair material for section restoration, is evaluated and residual bond strength is measured by considering unity with concrete. According to the evaluation of fire resistance performance of re-emulsification type polymer cement mortars, residual compressive strength was drastically decreased according to heating temperatures with an increase of polymer addition rate, and this seems to be attributable to the application of polymer film. In addition, an explosion phenomenon occurred frequently with an increase of addition rate, so this should be considered when selecting repair materials and processing.

• 한국산업융합학회 논문집
• /
• 제27권4_2호
• /
• pp.1001-1007
• /
• 2024

Recent deep learning algorithms for object detection in real-time play a crucial role in various applications such as autonomous driving, traffic monitoring, health care, and water quality monitoring. The size of small objects, in particular, significantly impacts the accuracy of detection models. However, data containing small objects can lead to underfitting issues in models. Therefore, this study developed a deep learning model capable of quickly detecting small objects to provide more accurate predictions. The RE-SOD (Residual block based Small Object Detector) developed in this research enhances the detection performance for small objects by using RGB separation preprocessing and residual blocks. The model achieved an accuracy of 1.0 in image classification and an mAP50-95 score of 0.944 in object detection. The performance of this model was validated by comparing it with real-time detection models such as YOLOv5, YOLOv7, and YOLOv8.

• 전자공학회논문지A
• /
• 제33A권10호
• /
• pp.1-11
• /
• 1996

In this paper, we evaluate the performance of DD-bardner (decision-directed gardner) algorithm. We derive an analytic gain of the thiming detector that is a function of SNR and an excess bandwidth, and verify the result by simulation. We also compare the DD-gardner algorithm with the bardner algorithm with respect to tracking performance and jitter performance under low SNR and a residual frequency component.

• Journal of Welding and Joining
• /
• 제25권1호
• /
• pp.42-48
• /
• 2007

For high performance and structural stability, application of high strength steel has continuously increased. However, the change of the base metal gives rise to problems with the accuracy management of the welded structure. It is attributed to the martensite phase transformation of the high strength low alloy steel weldment. The purpose of this study is to establish the predictive equation of transverse shrinkage and residual stress for the HY-100 weldment. In order to do it, high speed quenching dilatometer tests were performed to define a coefficient of thermal expansion(CTE) at the heating and cooling stage of HY-100 with various cooling rates. Uncoupled thermal-mechanical finite element(FE) models with CTE were proposed to evaluate the effect of the martensite phase transformation on transverse shrinkage and residual stresses at the weldment. FEA results were verified by comparing with experimental results. Based on the results of extensive FEA and experiments, the predictive equation of transverse shrinkage and longitudinal shrinkage force at the HY-100 weldment were formulated as the function of welding heat input/in-plane rigidity and welding heat input respectively.

• International Journal of Automotive Technology
• /
• 제8권6호
• /
• pp.697-704
• /
• 2007

Recently, a variable valve timing system has been widely adopted in internal combustion engine in order to improve the fuel economy and torque at low engine speed. In addition, it is known that varying valve timing according to the various engine operations could reduce exhaust gas, especially NOx, because of residual gas by valve overlap. In this study, to improve the low exhaust gas and fuel economy at part load condition, the residual gas and back flow of exhaust gas due to valve overlap were calculated computationally. Moreover, the characteristics of engine performances and NOx formations were investigated with the experiment of combination of intake and exhaust valve timing condition. Under these various valve operating conditions, the effects of both the positive valve overlap and negative valve overlap(valve underlap) were examined simultaneously. Finally, the characteristics of cyclic THC emission were analyzed by using Fast Response FID(FR-FID) in the cylinder, intake port and exhaust port positions. Besides, the effect of the different gradients of the valve timing change on engine performance was investigated and an optimum control strategy was suggested.

• Structural Engineering and Mechanics
• /
• 제53권5호
• /
• pp.867-880
• /
• 2015

Mechanical fastening is still one of the main methods used for joining components. Different techniques have been applied to reduce the effect of stress concentration of notches like fastener holes. In this work we evaluate the feasibility of combining laser shock peening (LSP) and cold expansion to improve fatigue crack initiation and propagation of open hole specimens made of 6061-T6 aluminum alloy. LSP is a new and competitive technique for strengthening metals, and like cold expansion, induces a compressive residual stress field that improves fatigue, wear and corrosion resistance. For LSP treatment, a Q-switched Nd:YAG laser with infrared radiation was used. Residual stress distribution as a function of depth was determined by the contour method. Compact tension specimens with a hole at the notch tip were subjected to LSP process and cold expansion and then tested under cyclic loading with R=0.1 generating fatigue cracks on the hole surface. Fatigue crack initiation and growth is analyzed and associated with the residual stress distribution generated by both treatments. It is observed that both methods are complementary; cold expansion increases fatigue crack initiation life, while LSP reduces fatigue crack growth rate.

• Earthquakes and Structures
• /
• 제11권4호
• /
• pp.649-664
• /
• 2016

A seismic damaged bridge may be hit again by a strong aftershock or another earthquake in a short interval before the repair work has been done. However, discussions about the impact of the unrepaired damages on the residual earthquake resistance of a steel bridge are very scarce at present. In this paper, nonlinear time-history analysis of a steel arch bridge was performed using multi-scale hybrid model. Two strong historical records of main shock-aftershock sequences were taken as the input ground motions during the dynamic analysis. The strain response, local deformation and the accumulation of plasticity of the bridge with and without unrepaired seismic damage were compared. Moreover, the effect of earthquake sequence on crack initiation caused by low-cycle fatigue of the steel bridge was investigated. The results show that seismic damage has little impact on the overall structural displacement response during the aftershock. The residual local deformation, strain response and the cumulative equivalent plastic strain are affected to some extent by the unrepaired damage. Low-cycle fatigue of the steel arch bridge is not induced by the earthquake sequences. Damage indexes of low-cycle fatigue predicted based on different theories are not exactly the same.

• 한국자동차공학회논문집
• /
• 제18권6호
• /
• pp.38-45
• /
• 2010

Welding structures are designed to endure its expected life. The most important factors are life. Especially on welded structure, fatigue strength is critical. So this study performed a research on Box and T shape weldment specimen to examine the influence of welding type. In this experiment, the results indicate Box shape was available in more than T shape. Fatigue tests were performed to evaluate the fatigue strength of the both as-welded and statically pre-loaded specimens by 3 point bending load. Fatigue life can be improved by using Ultrasonic Impact Treatment(UIT) effect. Ultrasonic Impact Treatment(UIT) is excellent for eliminating the tensile residual stresses and generating compressive residual stresses which elevate fatigue strength of welded structures. Also, this shows that welding part has better fatigue life and welding was performed well. In this study, to evaluate the Ultrasonic Impact Treatment(UIT) effect, for welding structure, the experiment was conducted at various levels of stress range between 100MPa and 500MPa. From the test results, it was indicated that fatigue performance was improving by Ultrasonic Impact Treatment(UIT)

• 대한기계학회논문집A
• /
• 제30권12호
• /
• pp.1543-1550
• /
• 2006

Since the performance of joints usually determines the structural efficiency of composite structures, an extensive knowledge of the behavior of adhesive joints and the related effect on joint strength is essential for design purposes. In this study, the torque capacity of adhesive joints was predicted using the combined thermal and mechanical analyses when the adherend was a composite tube. A finite element analysis was performed to evaluate residual thermal stresses developed in the joint, and mechanical s stresses in the adhesive were calculated including both the nonlinear adhesive behavior and the behavior of composite tubes. Three different joint failure modes were considered to predict joint failure: interfacial failure, adhesive bulk failure, and adherend failure. The influence of the composite adherend stacking angle on the residual thermal stresses was investigated, and how the residual thermal stresses affect the joint strength was also discussed. Finally, the predicted results were compared with experimental results available in literature.