• Journal of the Korea Society of Computer and Information
• /
• v.17 no.7
• /
• pp.45-52
• /
• 2012

In this paper, an crowd density in elevator estimation method based on multi-class Adaboost classifier is proposed. The SOM (Self-Organizing Map) based conventional methods have shown insufficient performance in practical scenarios and have weakness for low reproducibility. The proposed method estimates the crowd density using multi-class Adaboost classifier with texture features, namely, GLDM(Grey-Level Dependency Matrix) or GGDM(Grey-Gradient Dependency Matrix). In order to classify into multi-label, weak classifier which have better performance is generated by modifying a weight update equation of general Adaboost algorithm. The crowd density is classified into four categories depending on the number of persons in the crowd, which can be 0 person, 1-2 people, 3-4 people, and 5 or more people. The experimental results under indoor environment show the proposed method improves detection rate by about 20% compared to that of the conventional method.

• 제어로봇시스템학회:학술대회논문집
• /
• 1997.10a
• /
• pp.56-59
• /
• 1997

Vision system is mainly consist of three parts of perception, action. One of these parts, perception system detects visual target in surrounding environment. Block-based motion estimation with compensation is one of the popular approaches without accuracy. The hierarchical method the optical flow with gradient is used to improve optical flow time delay.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.3
• /
• pp.1121-1141
• /
• 2020

We propose a robust visual object tracking algorithm fusing a convolutional neural network tracker trained offline from a large number of video repositories and a color histogram based tracker to track objects for mixing immersive audio. Our algorithm addresses the problem of occlusion and large movements of the CNN based GOTURN generic object tracker. The key idea is the offline training of a binary classifier with the color histogram similarity values estimated via both trackers used in this method to opt appropriate tracker for target tracking and update both trackers with the predicted bounding box position of the target to continue tracking. Furthermore, a histogram similarity constraint is applied before updating the trackers to maximize the tracking accuracy. Finally, we compute the depth(z) of the target object by one of the prominent unsupervised monocular depth estimation algorithms to ensure the necessary 3D position of the tracked object to mix the immersive audio into that object. Our proposed algorithm demonstrates about 2% improved accuracy over the outperforming GOTURN algorithm in the existing VOT2014 tracking benchmark. Additionally, our tracker also works well to track multiple objects utilizing the concept of single object tracker but no demonstrations on any MOT benchmark.

• Journal of Biomedical Engineering Research
• /
• v.15 no.4
• /
• pp.481-488
• /
• 1994

In this paper a study of neuro-pathway estimation based on visual and somatosensory evoked potential is given. The evoked potentials which are caused by visual and somatosensory stimulation are detected by an average method. The forward problem that is estimating a scalp potential from a given electrical source in the brain is solved by using a triple concentric spherical shell model of the head and a single current dipole model of the neuron activity. The inverse problem which calculates a source position is solved by a least square fit between the model predicted potential and a given evoked potential measurement. The similarities between estimated sensory neuro-pathways and physiological brain function regions are verified.

• KIEAE Journal
• /
• v.7 no.4
• /
• pp.141-146
• /
• 2007

It is necessary to guarantee the safety, serviceability and durability of reinforced concrete structures over their service life. However, concrete structures represent a decrease in their durability due to the effects of external environments according to the passage of time, and such degradation in durability can cause structural degradation in materials. In concrete structures, some degradations in durability increase the corrosion of embedded rebars and also decrease the structural performance of materials. Thus, the structural condition assessment of RC materials damaged by corrosion of rebars becomes an important factor that judges needs to apply restoration. In order to detect the damage of reinforced concrete structures, a visual inspection, a nondestructive evaluation method(NDE) and a specific loading test have been employed. However, obscurities for visual inspection and inaccessible members raise difficulty in evaluating structure condition. For these reasons, detection of location and quantification of the damage in structures via structural response have been one of the very important topics in system identification research. The main objective of this project is to develope a methodologies for the damage identification via static responses of the members damaged by durability. Six reinforced concrete beams with variables of corrosion position and corrosion width were fabricated and the damage detections of corroded RC beams were performed by the optimization and the conjugate beam methods using static deflection. In results it is proved that the conjugate beam method could predict the damage of RC members practically.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.8
• /
• pp.1423-1443
• /
• 2011

Background compensation plays an important role in detecting and isolating object motion in visual tracking. Here, we propose a Genetic Hough Transform, which combines the Hough Transform and Genetic Algorithm, as a method for eliminating background motion. Our method can handle cases in which the background may contain only a few, if any, feature points. These points can be used to estimate the motion between two successive frames. In addition to dealing with featureless backgrounds, our method can successfully handle motion blur. Experimental comparisons of the results obtained using the proposed method with other methods show that the proposed approach yields a satisfactory estimate of background motion.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.20 no.4
• /
• pp.26-38
• /
• 1993

The purpose of this study was to suggest objective basic data for landscape research which cotained interdiciplinary approach between ecological and psychophysical method. For this, the naturalness and preference in greenary space, spatial image structure of physical elements have been analyzed by correlation coefficient and factor analysis algorithm. The results are as follows; 1. The relation between the estimation of naturalness and the preference in natural forest was deeply correlated. And the estimation of naturalness was higher than the estimation of DGN(Degree of Green Naturality). 2. The estimation of naturalnesss was decided by the physical features of forest and was different from the DGN. 3. Factors covering the spatial image of the forests have been found to be the 'overall evaluation', 'ability', 'naturalness' and 'idiosyncracy'. By using the control method for the number of factors, T.V. has been obtained as 61.68%. 4. The factors of the 'overall evauation', 'naturalness' were found to be the main factors determining the visual preference of greenary space.

• Journal of Korean Society of Forest Science
• /
• v.89 no.2
• /
• pp.145-160
• /
• 2000

This study is focused to the national park of Korean typical Sea Hally$\check{o}$ Haesang, and its visual resources and practiced inspect course by the way of suppositions and tests, to show the visual resource management objectively, and that of qualitative basic data. Accordingly by measuring the physical amount spatial structure with the visual amount originated from the Mesh Analyzing Method and the Visual Preference from the Scenic Beauty Estimation(S.B.E.) method and analyzed the valuation of the visual resource by Iverson method. Spatial image structure measured by Semantic Differential(S.D.) Scale was shown through the factor analysis algorithm for the analyzing psychological amount and examined the flowing out of decisive factor and the objective importance related to the mutual factors by appling the measurement of the visual quality. As a national Park, the visual factors that have natural landscape harmonized with forest, sky, surface of the water, curious stones and rocks, and temples should be escalated their values affirmatively so as to be the scenery of pointed direction and enjoyable, and it is of more needed for visual resource and its' controlling technique to make artificial structures more intentional planning and systemical setting. When we are viewing the improvement for the national park along with the visual resource management, reasonable level of development is needed, because when men interference surpass plantations and leasts will be damaged and the quality of natural landscape can be lowered, so it is needed to set up a management end, tangibly or clearly; and it is permitted limit coming and going ablably by accounting the suitable number for availing. But the controling end should be set in every level, positive management, very actively within the permissive varcability. It is the main business for the national park to prevent the damage from human for their gay life or to prevent the damage of a land carpet, and to restorate for the visual resource management.

• Journal of information and communication convergence engineering
• /
• v.12 no.1
• /
• pp.53-59
• /
• 2014

Motion estimation is a core part of most video compression systems since it directly affects the output video quality and the encoding time. The full search (FS) technique gives the highest visual quality but has the problem of a significant computational load. To solve this problem, we present in this paper a modified median (MMED) operation and advanced search strategies for fast motion estimation. The proposed MMED operation includes a temporally co-located motion vector (MV) to select an appropriate initial candidate. Moreover, we introduce a search procedure that reduces the number of thresholds and simplifies the early termination conditions for the determination of a final MV. The experimental results show that the proposed approach achieves substantial speedup compared with the conventional methods including the motion vector field adaptive search technique (MVFAST) and predictive MVFAST (PMVFAST). The proposed algorithm also improves the PSNR values by increasing the correlation between the MVs, compared with the FS method.

• Biomedical Science Letters
• /
• v.12 no.4
• /
• pp.439-442
• /
• 2006

To evaluate the postmortem changes in activities of oxygen free radical metabolizing enzymes, the rats were sacrificed with cervical dislocation and were kept in an incubator at $25^{\circ}C$, 70% of humidity for 12 hours. The activities of aniline hydroxylase, catalase, glutathione-S-transferase and superoxlde dismutase were decreased with the time. On the other hand, the activity and type conversion ratio (type D ${\to}$type O) of hepatic xanthine oxidase (XO) were gradually increased. From these changes of XO, the estimation of death time (mathematical equation) could be determined with the least square method. To clarify the cause of increasing XO activity, enzyme kinetics were examined. The Km values of XO were decreased with the time. In conclusion, the determination of liver XO activity might be used for the estimation of death time in the early postmortem period.