• Journal of Korea Multimedia Society
• /
• v.25 no.6
• /
• pp.807-818
• /
• 2022

In order to further improve the accuracy and time efficiency of behavior recognition in intelligent monitoring scenarios, a human behavior recognition algorithm based on YOLO combined with LSTM and CNN is proposed. Using the real-time nature of YOLO target detection, firstly, the specific behavior in the surveillance video is detected in real time, and the depth feature extraction is performed after obtaining the target size, location and other information; Then, remove noise data from irrelevant areas in the image; Finally, combined with LSTM modeling and processing time series, the final behavior discrimination is made for the behavior action sequence in the surveillance video. Experiments in the MSR and KTH datasets show that the average recognition rate of each behavior reaches 98.42% and 96.6%, and the average recognition speed reaches 210ms and 220ms. The method in this paper has a good effect on the intelligence behavior recognition.

• The Bulletin of The Korean Astronomical Society
• /
• v.44 no.2
• /
• pp.77.1-77.1
• /
• 2019

We present the results of the merger fraction evolution for galaxies in NEP-Wide field depending on star formation mode and redshift. We select the galaxies which have AKARI 9 ㎛ detections as a sample for large number of galaxies. We use multi-wavelength data from GALEX to Herschel, and Subaru HSC i-band images for analysis. We classify the merger galaxies with using Gini and M20, which are non-parametric calculated by statmorph code. We obtain the total infrared luminosity from the SED modeling with using one band, AKARI 9 ㎛. We find that the merger fractions of galaxies in all different star formation mode increase as the redshift increases. However, with fixed mass range of 10.5 < log(M🞵) < 11.5, the merger fractions of starbursts significantly increase as the redshift increases compared to those of main sequence and quiescent galaxies. We discuss the implications of these results in this poster.

• Journal of the Korea Concrete Institute
• /
• v.13 no.5
• /
• pp.457-465
• /
• 2001

Shrinkage strains of concrete slab in multi-story building are restrained by structural members such as columns or walls, then can induce cracks due to excessive shrinkage stress over tensile strength of member. In this study, a shrinkage stress analysis method of concrete slab in multi-story building considering not only material properties such as shrinkage, creep and reinforcement effect but also construction sequence is proposed. Tensile stresses of slab due to shrinkage are calculated by converting shrinkage strains into equivalent temperature gradients, creep that can release shrinkage stress can be considered by replacing the modulus of elasticity of concrete, Ec , to the effective secant modulus of elasticity of concrete, E$\_$eff/ Reinforcements are also considered by modeling them as equivalent beam elements in FEM program. Results of step by step analysis reflecting construction sequence summed up to calculate stresses of the whole building considering that shrinkage stresses of the building come from the difference of shrinkage between i-th floor and (i-1)-th floor, named as effecitive shrinkage, and it can be varied by construction sequence. The results of 10-story example building show that shrinkage stresses of lower floors are greater than those of upper floors, that is, stresses of lower floors(1∼2FI.) exceed modulus of rupture of concrete, but stress ratios of higher floors are in the range of 27.9∼92.8%.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.2 s.302
• /
• pp.113-120
• /
• 2005

3D reconstruction of a scene structure from un-calibrated image sequences has been long one of the central problems in computer vision. For 3D reconstruction in Euclidean space, projective reconstruction, which is classified into the merging method and the factorization, is needed as a preceding step. By calculating all camera projection matrices and structures at the same time, the factorization method suffers less from dia and error accumulation than the merging. However, the factorization is hard to analyze precisely long sequences because it is based on the assumption that all correspondences must remain in all views from the first frame to the last. This paper presents a new projective reconstruction method for recovery of 3D structure over long sequences. We break a full sequence into sub-sequences based on a quantitative measure considering the number of matching points between frames, the homography error, and the distribution of matching points on the frame. All of the projective reconstructions of sub-sequences are registered into the same coordinate frame for a complete description of the scene. no experimental results showed that the proposed method can recover more precise 3D structure than the merging method.

• Journal of Digital Convergence
• /
• v.15 no.12
• /
• pp.323-331
• /
• 2017

In this paper, four types of deposition control processes such as temperature, pressure, input/output(I/O), and gas were replaced by the Internet of Things(IoT) to analyze the data load and sequence procedure before and after the application of it. Through this analysis, we designed the load balancing in the sensing area of the deposition process by creating the sequence diagram of the deposition process. In order to do this, we were modeling of the sensor I/O according to the arrival process and derived the result of measuring the load of CPU and memory. As a result, it was confirmed that the reliability on the deposition processes were improved through performing some functions of the equipment controllers by the IoT. As confirmed through this paper, by applying the IoT to the deposition process, it is expected that the stability of the equipment will be improved by minimizing the load on the equipment controller even when the equipment is expanded.

• Journal of KIISE:Software and Applications
• /
• v.35 no.12
• /
• pp.780-788
• /
• 2008

Modeling hand poses and tracking its movement are one of the challenging problems in computer vision. There are two typical approaches for the reconstruction of hand poses in 3D, depending on the number of cameras from which images are captured. One is to capture images from multiple cameras or a stereo camera. The other is to capture images from a single camera. The former approach is relatively limited, because of the environmental constraints for setting up multiple cameras. In this paper we propose a method of reconstructing 3D hand poses from a 2D input image sequence captured from a single camera by means of Belief Propagation in a graphical model and recognizing a finger clicking motion using a hidden Markov model. We define a graphical model with hidden nodes representing joints of a hand, and observable nodes with the features extracted from a 2D input image sequence. To track hand poses in 3D, we use a Belief Propagation algorithm, which provides a robust and unified framework for inference in a graphical model. From the estimated 3D hand pose we extract the information for each finger's motion, which is then fed into a hidden Markov model. To recognize natural finger actions, we consider the movements of all the fingers to recognize a single finger's action. We applied the proposed method to a virtual keypad system and the result showed a high recognition rate of 94.66% with 300 test data.

• Journal of the Society of Naval Architects of Korea
• /
• v.38 no.2
• /
• pp.79-86
• /
• 2001

This study shows an attempt to generate an assembly sequence and its model for a ship block assembly using the graph theory and graph algorithms. To generate the ship block assembly, we propose four levels of the ship block assembly model such as "geometry mode1", "relational model", "sequential mode1", and "hierarchical model". To obtain the relational model, we used surface and surface intersection algorithm. The sequential model that represents a possible assembly sequence is made by using several graph algorithms from the relational model. The hierarchical model will be constructed from the sequential model in order to represent the block assembly tree and so forth. The purpose of the hierarchical model is to define an assembly tree and to generate the Bill Of Material(BOM). Lastly, the validity of the method proposed in this study is examined with application to ship block assembly models of a single type and double type according to four models mentioned above.

• Ocean and Polar Research
• /
• v.33 no.2
• /
• pp.159-169
• /
• 2011

Antifreeze proteins (AFPs) have ice binding affinity, depress freezing temperature and inhibit ice recystallization which protect cellular membranes in polar organisms. Recent structural studies of antifreeze proteins have significantly expanded our understanding of the structure-function relationship and ice crystal growth inhibition. Although AFPs (Type I-IV AFP from fish, insect AFP and Plant AFP) have completely different fold and no sequence homology, they share a common feature of their surface area for ice binding property. The conserved ice-binding sites are relatively flat and hydrophobic. For example, Type I AFP has an amphipathic, single ${\alpha}$-helix and has regularly spaced Thr-Ala residues which make direct interaction with oxygen atoms of ice crystals. Unlike Type I AFP, Type II and III AFP are compact globular proteins that contain a flat ice-binding patch on the surface. Type II and Type III AFP show a remarkable structural similarity with the sugar binding lectin protein and C-terminal domain of sialic acid synthase, respectively. Type IV is assumed to form a four-helix bundle which has sequence similarity with apolipoprotein. The results of our modeling suggest an ice-binding induced structural change of Type IV AFP. Insect AFP has ${\beta}$-helical structure with a regular array of Thr-X-Thr motif. Threonine residues of each Thr-X-Thr motif fit well into the ice crystal lattice and provide a good surface-surface complementarity. This review focuses on the structural characteristics and details of the ice-binding mechanism of antifreeze proteins.

• Communications of the Korean Mathematical Society
• /
• v.25 no.1
• /
• pp.119-128
• /
• 2010

Let ${X_n,\;n\geq1}$ be a sequence of independent identically distributed (i.i.d.) random variables (r.vs.), defined on a probability space ($\Omega$,A,P), and let ${N_n,\;n\geq1}$ be a sequence of positive integer-valued r.vs., defined on the same probability space ($\Omega$,A,P). Furthermore, we assume that the r.vs. $N_n$, $n\geq1$ are independent of all r.vs. $X_n$, $n\geq1$. In present paper we are interested in asymptotic behaviors of the random sum $S_{N_n}=X_1+X_2+\cdots+X_{N_n}$, $S_0=0$, where the r.vs. $N_n$, $n\geq1$ obey some defined probability laws. Since the appearance of the Robbins's results in 1948 ([8]), the random sums $S_{N_n}$ have been investigated in the theory probability and stochastic processes for quite some time (see [1], [4], [2], [3], [5]). Recently, the random sum approach is used in some applied problems of stochastic processes, stochastic modeling, random walk, queue theory, theory of network or theory of estimation (see [10], [12]). The main aim of this paper is to establish some results related to the asymptotic behaviors of the random sum $S_{N_n}$, in cases when the $N_n$, $n\geq1$ are assumed to follow concrete probability laws as Poisson, Bernoulli, binomial or geometry.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.383-388
• /
• 2005

Depth recovery in robot vision is an essential problem to infer the three dimensional geometry of scenes from a sequence of the two dimensional images. In the past, many studies have been proposed for the depth estimation such as stereopsis, motion parallax and blurring phenomena. Among cues for depth estimation, depth from lens translation is based on shape from motion by using feature points. This approach is derived from the correspondence of feature points detected in images and performs the depth estimation that uses information on the motion of feature points. The approaches using motion vectors suffer from the occlusion or missing part problem, and the image blur is ignored in the feature point detection. This paper presents a novel approach to the defocus technique based depth from lens translation using sequential SVD factorization. Solving such the problems requires modeling of mutual relationship between the light and optics until reaching the image plane. For this mutuality, we first discuss the optical properties of a camera system, because the image blur varies according to camera parameter settings. The camera system accounts for the camera model integrating a thin lens based camera model to explain the light and optical properties and a perspective projection camera model to explain the depth from lens translation. Then, depth from lens translation is proposed to use the feature points detected in edges of the image blur. The feature points contain the depth information derived from an amount of blur of width. The shape and motion can be estimated from the motion of feature points. This method uses the sequential SVD factorization to represent the orthogonal matrices that are singular value decomposition. Some experiments have been performed with a sequence of real and synthetic images comparing the presented method with the depth from lens translation. Experimental results have demonstrated the validity and shown the applicability of the proposed method to the depth estimation.