• 한국정보처리학회논문지
• /
• 제7권7호
• /
• pp.2254-2263
• /
• 2000

This paper proposes an efficient predictive image-compression technique based on vector quantization of blocks of pels. In the proposed method edge contents of blocks control the selection of predictors and block shapes as well. The maximum number of bits assigned to quantizers has been in creased to 3bits/pel from 1/5bits/pel, the setting employed by forerunners in predictive vector quantization of images. This increase prevents the saturation in SNR observed in their results in high bit rates. The variable block shape is instrumental in eh reconstruction of edges. The adaptive procedure is controlled by means of he standard deviation ofp rediction errors generated by a default predictor; the standard deviation address a decision table which can be set up beforehand. eh proposed method is characterized by overall improvements in image quality over A-VQ-PE and A-DCT VQ, both of which are known for their efficient use of vector quantizers.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1993년도 한국자동제어학술회의논문집(국내학술편); Seoul National University, Seoul; 20-22 Oct. 1993
• /
• pp.859-863
• /
• 1993

There are two methods to get 3-dimensional information by matching image pair feature-based matching and area-based matching. One of the problems in the area-based matching is how the optimal search region which gives accurate correlation between given point and its neighbors can be selected. In this paper, we proposed a new area-based matching algorithm which uses edge-features used in the conventional feature-based matching. It first selects matching candidates by feature-based and matches image pair with area-based method by taking these candidates as guidance to decision of search area. The results show that running time is reduced by optimizing search area(considering edge points and continuity of disparity), keeping on the precision as the conventional area-based matching method.

• IEIE Transactions on Smart Processing and Computing
• /
• 제2권4호
• /
• pp.189-196
• /
• 2013

This paper presents a two-stage approach to detect pedestrians in video sequences taken from a moving vehicle. The first stage is a preprocessing step, in which potential pedestrians are hypothesized. During the preprocessing step, a difference image is constructed using a global motion estimation, vertical and horizontal edge maps are extracted, and the color difference between the road and pedestrians are determined to create candidate regions where pedestrians may be present. The candidate regions are refined further using the vertical edge symmetry features of the pedestrians' legs. In the next stage, each hypothesis is verified using the integral channel features and an AdaBoost classifier. In this stage, a decision is made as to whether or not each candidate region contains a pedestrian. The proposed algorithm was tested on a range of dataset images and showed good performance.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1998년도 하계학술대회 논문집 G
• /
• pp.2222-2224
• /
• 1998

This paper presents a 3D feature extraction method using rough set theory. Using the stereo cameras, we obtain the raw images and then perform several processes including gradient computation and image matching process. Decision rule constructed via rough set theory determines whether a ceratin point in the image is 3D edge or not. We propose a method finding rules for 3D edge extraction using rough set.

• Smart Structures and Systems
• /
• 제32권4호
• /
• pp.219-233
• /
• 2023

Crack width is an important indicator to evaluate the health condition of the concrete structure. The crack width is measured by manual using crack width gauge commonly, which is time-consuming and laborious. In this paper, we have proposed a fast and simplified crack width quantification method via deep Q learning and geometric calculation. Firstly, the crack edge is extracted by using U-Net network and edge detection operator. Then, the intelligent decision of is made by the deep Q learning model. Further, the geometric calculation method based on endpoint and curvature extreme point detection is proposed. Finally, a case study is carried out to demonstrate the effectiveness of the proposed method, achieving high precision in the real crack width quantification.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1998년도 추계종합학술대회 논문집
• /
• pp.449-452
• /
• 1998

In this study, the basic image-board and algorithm has been developed to extract a road lane by modeling the driving process. The high speed processing enables an image capture, processing and prompt decision making. In order to high speed processing ASIC like FPGA was designed and integrated in one board system. The algorithm enabling road driving must recognize a straight and bend edge separately. The high speed image processing board using FPGA can be used in real-time decision makeing system for road driving and in the machine vision under bad working environments like a coal mine. And it also can be used in the safety control system in subway and in image input system of CCTV and CATV by designing the board to meet various user's needs.

• International Journal of Internet, Broadcasting and Communication
• /
• 제15권3호
• /
• pp.58-65
• /
• 2023

Bankruptcy is a significant risk for start-up companies, but with the help of cutting-edge artificial intelligence technology, we can now predict bankruptcy with detailed explanations. In this paper, we implemented the Category Boosting algorithm following data cleaning and editing using OpenRefine. We further explained our model using the Shapash library, incorporating domain knowledge. By leveraging the 5C's credit domain knowledge, financial analysts in banks or investors can utilize the detailed results provided by our model to enhance their decision-making processes, even without extensive knowledge about AI. This empowers investors to identify potential bankruptcy risks in their business models, enabling them to make necessary improvements or reconsider their ventures before proceeding. As a result, our model serves as a "glass-box" model, allowing end-users to understand which specific financial indicators contribute to the prediction of bankruptcy. This transparency enhances trust and provides valuable insights for decision-makers in mitigating bankruptcy risks.

• Journal of Information Processing Systems
• /
• 제20권2호
• /
• pp.226-238
• /
• 2024

Recently, multi-access edge computing (MEC) has emerged as a promising technology to alleviate the computing burden of vehicular terminals and efficiently facilitate vehicular applications. The vehicle can improve the quality of experience of applications by offloading their tasks to MEC servers. However, channel conditions are time-varying due to channel interference among vehicles, and path loss is time-varying due to the mobility of vehicles. The task arrival of vehicles is also stochastic. Therefore, it is difficult to determine an optimal offloading with resource allocation decision in the dynamic MEC system because offloading is affected by wireless data transmission. In this paper, we study computation offloading with resource allocation in the dynamic MEC system. The objective is to minimize power consumption and maximize throughput while meeting the delay constraints of tasks. Therefore, it allocates resources for local execution and transmission power for offloading. We define the problem as a Markov decision process, and propose an offloading method using deep reinforcement learning named deep deterministic policy gradient. Simulation shows that, compared with existing methods, the proposed method outperforms in terms of throughput and satisfaction of delay constraints.

• 스마트미디어저널
• /
• 제13권2호
• /
• pp.85-94
• /
• 2024

Artificial Intelligence of Things (AIoT), which combines AI and the Internet of Things (IoT), has recently gained popularity. Deep neural networks (DNNs) have achieved great success in many applications. Deploying complex AI models on embedded boards, nevertheless, may be challenging due to computational limitations or intelligent model complexity. This paper focuses on an AIoT-based system for smart sewing automation using edge devices. Our technique included developing a detection model and a decision tree for a sufficient testing scenario. YOLOv5 set the stage for our defective sewing stitches detection model, to detect anomalies and classify the sewing patterns. According to the experimental testing, the proposed approach achieved a perfect score with accuracy and F1score of 1.0, False Positive Rate (FPR), False Negative Rate (FNR) of 0, and a speed of 0.07 seconds with file size 2.43MB.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2008년도 하계종합학술대회
• /
• pp.849-850
• /
• 2008

Region adaptive demosaicing scheme based on the spectral spatial correlation is prosed in this paper. Region classification and adaptive demosaicing makes the decision of the edge direction more accurately.