• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• fall
• /
• pp.270-272
• /
• 2021

최근 영상신호처리에 대한 딥러닝 기술이 비약적으로 발전함에 따라 다양한 방면으로 시도되고 있다. 그 중 machine level vision 에서 인지 기능을 하는 optical flow 를 end-to-end 학습 방식으로 제시하여 고성능 결과물을 도출하는 RAFT(Recurrent All-pairs Field Transform for Optical flow, 2020)에 대해 분석하고자 한다. RAFT 는 입력된 두 이미지에 대한 4D correlation volume 을 구축하여 모든 픽셀에 대한 정보를 사용한다. 또한, recurrent neural network 에서 차용한 반복적인 연산 학습 구조를 통하여 결과물인 flow field 의 정확도를 높인다. 해당 모델은 stereo dataset 을 사용하는 다른 모델에 비해 학습 시간이 짧고 용량이 작으면서 error rate 은 낮은 모습을 보인다. 현재 많은 연구에서 optical flow 를 접목하려는 움직임을 보이고 있고 다양하게 활용될 가능성이 다분하다는 점에서 주목할 가치가 있다.

• Journal of Mechanical Science and Technology
• /
• v.15 no.3
• /
• pp.300-308
• /
• 2001

Tracking of moving objects within video streams is a complex and time-consuming process. Large number of moving objects increases the time for computation of tracking the moving objects. Because of large computations, there are real-time processing problems in tracking of moving objects. Also, the change of environment causes errors in estimation of tracking information. In this paper, we present a new method for tracking of moving objects using optical flow motion analysis. Optical flow represents an important family of visual information processing techniques in computer vision. Segmenting an optical flow field into coherent motion groups and estimating each underlying motion are very challenging tasks when the optical flow field is projected from a scene of several moving objects independently. The problem is further complicated if the optical flow data are noisy and partially incorrect. Optical flow estimation based on regulation method is an iterative method, which is very sensitive to the noisy data. So we used the Combinatorial Hough Transform (CHT) and Voting Accumulation for finding the optimal constraint lines. To decrease the operation time, we used logical operations. Optical flow vectors of moving objects are extracted, and the moving information of objects is computed from the extracted optical flow vectors. The simulation results on the noisy test images show that the proposed method finds better flow vectors and more correctly estimates the moving information of objects in the real time video streams.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.8
• /
• pp.132-139
• /
• 2015

Visual SLAM(Simultaneous Localization And Mapping) has been used widely to estimate a mobile robot's location. Visual SLAM estimates relative motions with static visual features over image sequence. Because visual SLAM methods assume generally static features in the environment, we cannot obtain precise results in dynamic situation including many moving objects: cars and human beings. This paper presents a stereo vision based SLAM method in dynamic environment. First, we extract disparity map with stereo vision and compute optical flow. We then compute disparity change that is the estimated flow field between stereo views. After examining the disparity change value, we detect ROIs(Region Of Interest) in disparity space to determine dynamic scene objects. In indoor environment, many structural planes like walls may be determined as false dynamic elements. To solve this problem, we segment the scene into planar structure. More specifically, disparity values by the stereo vision are projected to X-Z plane and we employ Hough transform to determine planes. In final step, we remove ROIs nearby the walls and discriminate static scene elements in indoor environment. The experimental results show that the proposed method can obtain stable performance in dynamic environment.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.3
• /
• pp.107-112
• /
• 2018

In this paper, optical flow based keypoint detection and tracking technique is proposed for the collaboration between flying drone with vision system and ground robots. There are many challenging problems in target detection research using moving vision system, so we combined the improved FAST algorithm and Lucas-Kanade method for adopting the better techniques in each feature detection and optical flow motion tracking, which results in 40% higher in processing speed than previous works. Also, proposed image binarization method which is appropriate for the given marker helped to improve the marker detection accuracy. We also studied how to optimize the embedded system which is operating complex computations for intelligent functions in a very limited resources while maintaining the drone's present weight and moving speed. In a future works, we are aiming to develop collaborating smarter robots by using the techniques of learning and recognizing targets even in a complex background.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.8
• /
• pp.2948-2963
• /
• 2015

In the field of computer vision, visual surveillance systems have recently become an important research topic. Growth in this area is being driven by both the increase in the availability of inexpensive computing devices and image sensors as well as the general inefficiency of manual surveillance and monitoring. In particular, the ultimate goal for many visual surveillance systems is to provide automatic activity recognition for events at a given site. A higher level of understanding of these activities requires certain lower-level computer vision tasks to be performed. So in this paper, we propose an intelligent activity recognition model that uses a structure learning method and a classification method. The structure learning method is provided as a K2-learning algorithm that generates Bayesian networks of causal relationships between sensors for a given activity. The statistical characteristics of the sensor values and the topological characteristics of the generated graphs are learned for each activity, and then a neural network is designed to classify the current activity according to the features extracted from the multiple sensor values that have been collected. Finally, the proposed method is implemented and tested by using PETS2013 benchmark data.

• Journal of the Korea Society of Computer and Information
• /
• v.11 no.5 s.43
• /
• pp.329-336
• /
• 2006

This paper describes a lane change assistance system for the help of safe lane change, which detects vehicles approaching from the rear side by using a computer vision algorithm and notifies the possibility of safe lane change to a driver. In case a driver tries to lane change, the proposed system can detect vehicles and keep track of them. After detecting side lane lines, region of interest for vehicle detection is decided. For detection a vehicle, optical flow technique is applied. The experimental result of the proposed algorithm and system showed that the vehicle detection rate was 91% and the embedded system would have application to a lane change assistance system being commercialized in the near future.

• 한국전산유체공학회:학술대회논문집
• /
• 2008.03b
• /
• pp.498-501
• /
• 2008

Knowledge of airflow characteristics in nasal cavities is essential to understand the physiological and pathological aspects of nasal breathing. Several studies have utilized physical models of the healthy nasal cavity to investigate the relationship between nasal anatomy and airflow. In our laboratory, there have been a series of experimental investigations on the nasal airflow in normal, abnormal, and deformed nasal cavity models cavity models by PIV under both constant and periodic flow conditions. In this time normal and several deformed nasal cavity models, which simulate surgical operation, Turbinectomy, are investigated numerically by the FVM general purpose code and PIV analysis. The comparisons of these results are appreciated. Dense CT data and careful treatment of model surface under the ENT doctor's advice provide more sophisticated cavity models. The Davis (LaVision Co.) code is used for PIV flow analysis. Average and RMS distributions have been obtained for inspirational and expirational nasal airflows in the normal and deformed nasal cavities.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.2 no.2
• /
• pp.120-133
• /
• 2008

This paper presents a novel approach for facial motion tracking and facial expression cloning to create a realistic facial animation of a 3D avatar. The exact head pose estimation and facial expression tracking are critical issues that must be solved when developing vision-based computer animation. In this paper, we deal with these two problems. The proposed approach consists of two phases: dynamic head pose estimation and facial expression cloning. The dynamic head pose estimation can robustly estimate a 3D head pose from input video images. Given an initial reference template of a face image and the corresponding 3D head pose, the full head motion is recovered by projecting a cylindrical head model onto the face image. It is possible to recover the head pose regardless of light variations and self-occlusion by updating the template dynamically. In the phase of synthesizing the facial expression, the variations of the major facial feature points of the face images are tracked by using optical flow and the variations are retargeted to the 3D face model. At the same time, we exploit the RBF (Radial Basis Function) to deform the local area of the face model around the major feature points. Consequently, facial expression synthesis is done by directly tracking the variations of the major feature points and indirectly estimating the variations of the regional feature points. From the experiments, we can prove that the proposed vision-based facial expression cloning method automatically estimates the 3D head pose and produces realistic 3D facial expressions in real time.

• Journal of Forest and Environmental Science
• /
• v.28 no.1
• /
• pp.1-7
• /
• 2012

Tropical forests whether fragmented or undisturbed or be they equatorial or deciduous, remain the storehouse of biodiversity for hundreds of thousands of plant and animal species. This unique characteristic continues to attract a wide range of scientists and international organizations to study and attempt to understand tropical forest ecosystems. Gene flow is mediated by pollen, seed and seedling dispersal, and factors affecting this gene flow include phenology, spatial distribution, population structures, seed predation, sexual and mating systems as well as physical and biological barriers to gene flow. Two methods are used in measuring gene flow: direct method that relies on the actual observation of seed and pollen dispersal, whereas indirect methods involve the use of genetic markers such as allozymes and DNA techniques. Political strife, extreme natural and artificial disasters, the lack of a comprehensive forestry research vision, coupled with difficult socio-economic conditions in Africa have made the environment quite difficult for sustained research activities on the part of those undertaking or wishing to undertake such studies. Gene flow studies in this region are few and far between. This review elaborates on the mechanisms of gene flow mediation in Sub-Saharan Africa.

• Journal of Digital Convergence
• /
• v.16 no.12
• /
• pp.373-381
• /
• 2018

This study investigates convergent influence on TOEIC learning flow and its association with self efficacy, academic control and school resilience among health college students. Data collection was carried out using a self-administered questionnaire from May 1 to May 29, 2018 and the target was randomly selected 255 health college students in TOEIC class from college located in J city. TOEIC learning flow was positively correlated with self efficacy, academic control and school resilience. With the analysis of covariance structure, we could confirm relationship among self efficacy, academic control, school resilience and TOEIC learning flow. School resilience was more influential on TOEIC learning flow than self efficacy and academic control. The results of the study indicate that the efforts to manage these factors are required to increase TOEIC learning flow of health college students in TOEIC class. The results are expected to be used to develop TOEIC learning curriculum increasing the TOEIC learning flow among health college students in TOEIC class. In the following study, the survey about additional factors of influence on TOEIC learning flow will be needed.