• Journal of the Korea Society of Computer and Information
• /
• v.27 no.12
• /
• pp.59-68
• /
• 2022

This paper presents a method for 1:1 verification by comparing the similarity between the given real product image and the drawing image. The proposed method combines two existing CNN-based deep learning models to construct a Siamese Network. After extracting the feature vector of the image through the FC (Fully Connected) Layer of each network and comparing the similarity, if the real product image and the drawing image (front view, left and right side view, top view, etc) are the same product, the similarity is set to 1 for learning and, if it is a different product, the similarity is set to 0. The test (inference) model is a deep learning model that queries the real product image and the drawing image in pairs to determine whether the pair is the same product or not. In the proposed model, through a comparison of the similarity between the real product image and the drawing image, if the similarity is greater than or equal to a threshold value (Threshold: 0.5), it is determined that the product is the same, and if it is less than or equal to, it is determined that the product is a different product. The proposed model showed an accuracy of about 71.8% for a query to a product (positive: positive) with the same drawing as the real product, and an accuracy of about 83.1% for a query to a different product (positive: negative). In the future, we plan to conduct a study to improve the matching accuracy between the real product image and the drawing image by combining the parameter optimization study with the proposed model and adding processes such as data purification.

• The KIPS Transactions:PartB
• /
• v.9B no.3
• /
• pp.359-368
• /
• 2002

In this paper, we propose effective boundary line extraction algorithm for moving objects by matching error image and moving vectors, and fast tracking algorithm for moving object by partial boundary lines. We extracted boundary line for moving object by generating seeds with probability distribution function based on Watershed algorithm, and by extracting boundary line for moving objects through extending seeds, and then by using moving vectors. We processed tracking algorithm for moving object by using a part of boundary lines as features. We set up a part of every-direction boundary line for moving object as the initial feature vectors for moving objects. Then, we tracked moving object within current frames by using feature vector for the previous frames. As the result of the simulation for tracking moving object on the real images, we found that tracking processing of the proposed algorithm was simple due to tracking boundary line only for moving object as a feature, in contrast to the traditional tracking algorithm for active contour line that have varying processing cost with the length of boundary line. The operations was reduced about 39% as contrasted with the full search BMA. Tracking error was less than 4 pixel when the feature vector was $(15\times{5)}$ through the information of every-direction boundary line. The proposed algorithm just needed 200 times of search operation.

• Journal of KIISE
• /
• v.44 no.8
• /
• pp.813-821
• /
• 2017

Entity Linking find the meaning of an entity mention, which indicate the entity using different expressions, in a user's query by linking the entity mention and the entity in the knowledge base. This task has four challenges, including the difficult knowledge base construction problem, multiple presentation of the entity mention, ambiguity of entity linking, and NIL entity recognition. In this paper, we first construct the entity name dictionary based on Wikipedia to build a knowledge base and solve the multiple presentation problem. We then propose various methods for NIL entity recognition and solve the ambiguity of entity linking by training the support vector machine based on several features, including the similarity of the context, semantic relevance, clue word score, named entity type similarity of the mansion, entity name matching score, and object popularity score. We sequentially use the proposed two methods based on the constructed knowledge base, to obtain the good performance in the entity linking. In the result of the experiment, our system achieved 83.66% and 90.81% F1 score, which is the performance of the NIL entity recognition to solve the ambiguity of the entity linking.

• The Journal of the Korea Contents Association
• /
• v.7 no.2
• /
• pp.11-20
• /
• 2007

AAM has been successfully applied to face alignment, but its performance is very sensitive to initial values. In this paper, we propose a face alignment method using progressive AAM. The proposed method consists of two stages; modelling and relation derivation stage and fitting stage. Modelling and relation derivation stage first builds two AAM models; the inner face AAM model and the whole face AAM model and then derive the relation matrix between the inner face AAM model parameter vector and the whole face AAM model parameter vector. The fitting stage is processed progressively in two phases. In the first phase, the proposed method finds the feature parameters for the inner facial feature points of a new face, and then in the second phase it localizes the whole facial feature points of the new face using the initial values estimated utilizing the inner feature parameters obtained in the first phase and the relation matrix obtained in the first stage. Through experiments, it is verified that the proposed progressive AAM-based face alignment method is more robust with respect to pose, and face background than the conventional basic AAM-based face alignment.

• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.4
• /
• pp.383-391
• /
• 2010

A Lifelog management system provides users with services to store, manage, and search their life logs. This paper proposes a fully-automatic collecting method of real world social contacts and lifelog search engine using collected social contact information as keyword. Wireless short-distance network devices in mobile phones are used to detect social contacts of their users. Human-Bluetooth relationship matrix is built based on the frequency of a human-being and a Bluetooth device being observed at the same time. Results show that with 20% of social contact information out of full social contact information of the observation times used for calculation, 90% of human-Bluetooth relationship can be correctly acquired. A lifelog search-engine that takes human names as keyword is suggested which compares two vectors, a row of Human-Bluetooth matrix and a vector of Bluetooth list scanned while a lifelog was created, using vector information retrieval model. This search engine returns more lifelog than existing text-matching search engine and ranks the result unlike existing search-engine.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.34 no.12
• /
• pp.1087-1092
• /
• 2010

A micro 3D-PTV system has been constructed using a single camera system. Two viewing holes were created behind the object lens of the microscopic system to construct a stereoscopic viewing image. A hybrid recursive PTV algorithm was used. A concept of epipolar line was adopted to eliminate many spurious candidates. Three-dimensional velocity vector fields were obtained by calculating the three-dimensional displacements of particles that were identified as being identical. The system consists of a laser light source (Ar-ion, 500 mW), one high-definition camera ($1028{\times}1024$ pixels, 500 fps), a circular plate with two viewing holes, and a host computer. The performance of the developed algorithm was tested using artificial images. The characteristic of the vector recovery ratio was investigated for the particle numbers. A micro backward-facing step channel ($H{\times}h{\times}W:\;36{\mu}m{\times}70{\mu}m{\times}3000{\mu}m$) was measured using the developed measurement system. The results were in good qualitative agreement with other results.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.4B
• /
• pp.771-782
• /
• 2000

In this paper, an H.263 video codec is implemented by adopting the concept of hardware and software co-design. Each module of the codec is investigated to find which approach between hardware and software is better to achieve real-time processing speed as well as flexibility. The hardware portion includes motion-related engines, such as motion estimation and compensation, and a memory control part. The remaining portion of theH.263 video codec is implemented in software using a RISC processor. This paper also introduces efficient design methods for hardware and software modules. In hardware, an area-efficient architecture for the motion estimator of a multi-resolution block matching algorithm using multiple candidates and spatial correlation in motion vector fields (MRMCS), is suggested to reduce the chip size. Software optimization techniques are also explored by using the statistics of transformed coefficients and the minimum sum of absolute difference (SAD)obtained from the motion estimator.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.43 no.6 s.312
• /
• pp.57-64
• /
• 2006

Motion estimation plays an important role for the compression of video signals. The proposed method utilizes an adaptive threshold and characteristics of a distribution of SAD (sum of absolute difference). Generally, the more complex the SAD distribution is, the larger SAD value tends to be. This proposed algorithm tries to reduce the search points in a simple distribution but increase them in a complex distribution to avoid local minima. A macro block is divided into 9 areas. One of them chosen using spatio-temporal correlation is called the primary area and the others are called the secondary area that will be searched to avoid local minima. The proposed algorithm decides if just one area (the primary area or the secondary area) will be enough to be searched or both areas should be searched, using adaptive threshold. Compared with famous motion estimation algorithms, the simulation result shows that the searching points per macro block and MSE decreases about 16.4% and 32.83 respectively on the average.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.8
• /
• pp.3551-3567
• /
• 2016

The Low Earth Orbit satellite network has the unique characteristics of the non-uniform and time-variant traffic load distribution, which often causes severe link congestion and thus results in poor performance for delay-sensitive flows, especially when the network is heavily loaded. To solve this problem, a novel adaptive routing algorithm, referred to as the delay-oriented adaptive routing algorithm (DOAR), is proposed. Different from current reactive schemes, DOAR employs Destination-Sequenced Distance-Vector (DSDV) routing algorithm, which is a proactive scheme. DSDV is extended to a multipath QoS version to generate alternative routes in active with real-time delay metric, which leads to two significant advantages. First, the flows can be timely and accurately detected for route adjustment. Second, it enables fast, flexible, and optimized QoS matching between the alternative routes and adjustment requiring flows and meanwhile avoids delay growth caused by increased hop number and diffused congestion range. In addition, a retrospective route adjustment requesting scheme is designed in DOAR to enlarge the alternative routes set in the severe congestion state in a large area. Simulation result suggests that DOAR performs better than typical adaptive routing algorithms in terms of the throughput and the delay in a variety of traffic intensity.

• Journal of Institute of Control, Robotics and Systems
• /
• v.20 no.12
• /
• pp.1259-1266
• /
• 2014

Road signs provide important safety information about road and traffic conditions to drivers. Road signs include not only common traffic signs but also warning information regarding unexpected obstacles and road constructions. Therefore, accurate detection and identification of road signs is one of the most important research topics related to safe driving. In this paper, we propose a 3-D vision technique to automatically detect and track road signs in a video sequence which is acquired from a stereo vision camera mounted on a vehicle. First, color information is used to initially detect the sign candidates. Second, the SVM (Support Vector Machine) is employed to determine true signs from the candidates. Once a road sign is detected in a video frame, it is continuously tracked from the next frame until it is disappeared. The 2-D position of a detected sign in the next frame is predicted by the 3-D motion of the vehicle. Here, the 3-D vehicle motion is acquired by using the 3-D pose information of the detected sign. Finally, the predicted 2-D position is corrected by template-matching of the scaled template of the detected sign within a window area around the predicted position. Experimental results show that the proposed method can detect and track many types of road signs successfully. Tracking comparisons with two different methods are shown.