• Journal of Intelligence and Information Systems
• /
• v.15 no.2
• /
• pp.33-48
• /
• 2009

Capsule endoscopy is one of the most remarkable inventions in last ten years. Causing less pain for patients, diagnosis for entire digestive system has been considered as a most convenience method over a normal endoscope. However, it is known that the diagnosis process typically requires very long inspection time for clinical experts because of considerably many duplicate images of same areas in human digestive system due to uncontrollable movement of a capsule endoscope. In this paper, we propose a method for clinical diagnosticians to get highly valuable information from capsule-endoscopy video. Our software system consists of three global maps, such as movement map, characteristic map, and brightness map, in temporal domain for entire sequence of the input video. The movement map can be used for effectively removing duplicated adjacent images. The characteristic and brightness maps provide frame content analyses that can be quickly used for segmenting regions or locating some features(such as blood) in the stream. Our experiments show the results of four patients having different health conditions. The result maps clearly capture the movements and characteristics from the image frames. Our method may help the diagnosticians quickly search the locations of lesion, bleeding, or some other interesting areas.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.45 no.7
• /
• pp.32-37
• /
• 2008

Satellite communication systems including the DVB-S2 (Digital Video Broadcasting - Satellite Version 2) system require operations under low signal-to-noise ratio (SNR) and large frequency offset values, and the initial frame synchronization process necessitates a robust correlation method. While a variety of conventional correlation structures exist for the initial synchronization, each method has different characteristics and performance in different channel environments. In this paper, we propose new correlation methods which exhibit enhanced performance in low SNR and large frequency offsets, and analyze their performance. The proposed methods use the magnitude sum and vector sum of extended differential correlation values, to maximize the correlation between the received signal and the synchronization sequence by using the spanned differential correlation result. The magnitude sum correlation method has better performance compared to conventional methods including the approximated ML (Maximum likelihood) method for SNR values below 4 dB with or without frequency offsets. The vector sum correlation method has improved performance over the magnitude sum method for channels with relatively small frequency offsets.

• Smart Media Journal
• /
• v.13 no.2
• /
• pp.68-84
• /
• 2024

This paper describes the development of a system and algorithms for high-quality welfare services by recognizing behavior development indicators (activity, sociability, danger) in children aged 0 to 2 years old using action recognition technology. Action recognition targeted 11 behaviors from lying down in 0-year-olds to jumping in 2-year-olds, using data directly obtained from actual videos provided for research purposes by three nurseries in the Gwangju and Jeonnam regions. A dataset of 1,867 actions from 425 clip videos was built for these 11 behaviors, achieving an average recognition accuracy of 97.4%. Additionally, for real-world application, the Edge Video Analyzer (EVA), a behavior analysis device, was developed and implemented with a region-specific random frame selection-based PoseC3D algorithm, capable of recognizing actions in real-time for up to 30 people in four-channel videos. The developed system was installed in three nurseries, tested by ten childcare teachers over a month, and evaluated through surveys, resulting in a perceived accuracy of 91 points and a service satisfaction score of 94 points.

• Journal of Broadcast Engineering
• /
• v.12 no.2
• /
• pp.93-101
• /
• 2007

In heterogeneous networks SVC (Scalabile Video Coding) will be used for IPTV service. This paper analyses how to determine the optimal inter-layer reference scheme according to final level to be displayed in hybrid scalable coding which consists of spatial, quality and temporal layer. It determines where to stop layering quality layer stacks in lower spatial layer according to the relationship between noise induced by loss of high frequency component eliminated by filter in order to get rid of aliasing when spatial layering is processed and noise induced by quantization when quality layering is processed. This paper shows the choice of the level of layering between spatial and quality to get better coding efficiency and then presents what is needed for determining it.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.40 no.5
• /
• pp.347-356
• /
• 2003

Camera pose and scene geometry estimation from video sequences is widely used in various areas such as image composition. Structure and motion recovery based on the auto calibration algorithm can insert synthetic 3D objects in real but un modeled scenes and create their views from the camera positions. However, most previous methods require bundle adjustment or non linear minimization process [or more precise results. This paper presents a new auto' calibration algorithm for video sequence based on two steps: the one is key frame selection, and the other removes the key frame with inaccurate camera matrix based on an absolute quadric estimation by LMedS. In the experimental results, we have demonstrated that the proposed method can achieve a precise camera pose estimation and scene geometry recovery without bundle adjustment. In addition, virtual objects have been inserted in the real images by using the camera trajectories.

• Korean Journal of Applied Biomechanics
• /
• v.33 no.4
• /
• pp.147-154
• /
• 2023

Objective: This study aimed 1) to compare the Landing Error Scoring System (LESS) score and movement patterns during landing of the lesser dorsiflexion range of motion (LDFROM) group to that with the greater dorsiflexion range of motion group, and 2) to identify the correlation between the weight-bearing dorsiflexion range of motion (WBDF ROM), LESS score, and movement patterns during landing. Method: Fifty health adults participated in this study. WBDF ROM was measured using the weight bearing lunge test while movement patterns during landing was assessed using the LESS. The joint angles of the ankle, knee and hip joints during landing were analyzed using the 2D video analysis. After mean value of WBDF ROM was calculated, participants were divided into two groups (GDFROM and LDFROM) based on the mean value. The Mann-Whiteny 𝒰 test was used to identify differences in movement strategies during landing between two groups and the Pearson's correlation analysis was performed to determine relationships between WBDF ROM and movement strategies. Results: The LDFROM group showed the poorer LESS score and stiffer landing kinematics during landing compared to the GDFROM group (p<0.05). In addition, DFROM was significantly related to the LESS score and landing kinematics (p<0.05) except for total hip excursion (p=0.228). Conclusion: Our main findings showed that the LDFROM group had poorer landing quality and stiffer landing movements compared to the GDFROM group. In addition, increase of WBDF ROM significantly improved landing quality and soft-landing movements. To reduce shock during landing such as ground reaction forces, individuals need to better utilize WBDF ROM and lower extremity movements based on our findings. Therefore, intervention programs for safer landings should include exercises that increase WBDF ROM and utilize eccentric contraction.

• Smart Media Journal
• /
• v.12 no.11
• /
• pp.48-56
• /
• 2023

The technology of Three-dimensional human posture estimation is used in sports, motion recognition, and special effects of video media. Among various methods for this, multi-view 3D human pose estimation is essential for precise estimation even in complex real-world environments. But Existing models for multi-view 3D human posture estimation have the disadvantage of high order of time complexity as they use 3D feature maps. This paper proposes a method to extend an existing monocular viewpoint multi-frame model based on Transformer with lower time complexity to 3D human posture estimation for multi-viewpoints. To expand to multi-viewpoints our proposed method first generates an 8-dimensional joint coordinate that connects 2-dimensional joint coordinates for 17 joints at 4-vieiwpoints acquired using the 2-dimensional human posture detector, CPN(Cascaded Pyramid Network). This paper then converts them into 17×32 data with patch embedding, and enters the data into a transformer model, finally. Consequently, the MLP(Multi-Layer Perceptron) block that outputs the 3D-human posture simultaneously updates the 3D human posture estimation for 4-viewpoints at every iteration. Compared to Zheng[5]'s method the number of model parameters of the proposed method was 48.9%, MPJPE(Mean Per Joint Position Error) was reduced by 20.6 mm (43.8%) and the average learning time per epoch was more than 20 times faster.

• Archives of design research
• /
• v.18 no.3 s.61
• /
• pp.209-218
• /
• 2005

Consciously or not, producers have to make many aesthetic choices in creative process of video production. If there are general acceptable aesthetic principles to make right choice it would be guideline of aesthetic decision to somewhat reduce mistakes and errors in the process. This paper proposes a theoretical approach on establishing the media aesthetic principle of 3D animation camera working, which is the most suitable for animation production context. We describe the Herbert Zettl's applied media aesthetics related directly to the camera, which is about the two-Dimensional field focusing on aspect radio and forces within the screen, three-dimensional field focusing on depth, volume, and four-dimensional field focusing on time and motion. In order to have theoretical approach we made an analysis on comparing a camera working of movie with 3D computer animation's one, and reconstructed these basic principles to be suited for the 3D animation production. When applied media aesthetics of the traditional camera working are applied to the 3D animation production, it could be an efficient guideline for it. Futhermore, if we develop the research for the relationship with various visual languages with the basis of these principles, the theory of creative picture composition method for the 3D animation production will be logically and systematically established.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2013.10a
• /
• pp.919-921
• /
• 2013

영상시장의 개척과 디지털 기술의 발전과 더불어 차세대 3D 입체영상기술에 대한 관심과 수요가 증가하고 있다. 입체 정보는 크게 '단안 입체 정보(monoscopic depth cue)'와 '양안 입체 정보(stereoscopic depth cue)'로 분류 할 수 있다. 단안 입체 정보는 은폐, 상대적 크기, 상대적 밀도, 시야 안의 높이, 공기투시, 운동투시, 초점조절인 7가지로 경험에 의한 입체감을 지각하는 것을 말하며 양안 입체 정보는 두 눈으로 볼 때 처음으로 깊이를 지각 할 수 있는 것으로 크게 '동시시(simultaneous perception)', '융합(sensory fusion)', '입체시(stereoscopic vision)'의 세종류의 기능으로 분류한다. 3D 촬영은 이 양안시의 원리를 이용하여 두 대의 카메라의 좌우 영상을 합성하여 깊이감 있는 영상을 만들어 내게 된다. 본 논문에서는 3D 촬영방법은 촬영방식에 따라 크게 평행방식, 직교방식, 교차방식이 있는데 이중 중 원거리 촬영에 유리한 교차방식을 활용하여 사이드 바이 사이드 리그(Rig; 카메라를 수평으로 설치할 수 있도록 만들어진 장치)를 원거리 촬영에 맞게 축간거리를 기존의 리그 사이즈보다 2배 이상 긴 리그를 제작하여 보다 먼 거리에서의 상이한 좌우 영상획득이 가능하도록 설계하였다. 또한, 일정한 간격에 따라 피사체를 촬영하면서 거리에 따른 양 카메라의 가장 이상적인 IOD(Interocular Distance)와 CONV(Convergence)를 찾고, 교차방식촬영에 따른 특징적인 아티팩트인 키스톤 왜곡(Keystone distance)의 보정을 통한 원거리 입체영상을 효과적으로 획득하는데 본 연구방법을 제안하고자 한다.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2005.03a
• /
• pp.401-408
• /
• 2005

Recently shield TBM tunnellings are being applied to subway construction in Korean cities. Generally these kinds of tunnellings have the problems in the stability of ground such as subsidence because urban subway is constructed in the shallow depth. A sinkhole occurred on the road just above the tunnel during tunneling in Kwangju, so a survey for upper layer of the tunnel was needed. But conventional Ground Probing Radar can't be applicable due to the presence of steel-mesh screen in the shield segment, so no existent geophysical method is applicable in this site. Because the outer surface of each shield segment is electrically insulated, dipole-dipole resistivity method which is popular in engineering site investigation, was tried to this survey for the first time. Specially manufactured flexible ring-type electrodes were installed into the grouting holes at an interval of 2.4 m on the ceiling. The K-Ohm II system which has been developed by KIGAM and tested successfully in many sites, was used in this site. The system consists of 1000Volt-1Ampere constant-current transmitter, optically isolated 24 bit sigma-delta A/D conversion receiver - maximum 12 channel simultaneous measurements, and graphical automatic acquisition software for easy data quality check in real time. Borehole camera logging with circular white LED lighting was also done to investigate the state of the layer. Measured resistivity data lack of some stations due to failing opening lids of holes, shows general high-low trend well. The dipole-dipole resistivity inversion results discriminate (1) one approximately 4 meter diameter cavity (grouted but incompletely hardened, so low resistivity - less than $30{\Omega}m$), (2) weak zone (100-200${\Omega}m$), and (3) hard zone (high resistivity - more than 1000${\Omega}m$) very well for the distance of 320 meters. The 2-D inversion neglects slight absolute 3-D effect, but we can get satisfactory and useful information. Acquired resistivity section and video tapes by borehole camera logging will be reserved and reused if some problem occurs in this site in the future.