• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.05a
• /
• pp.519-521
• /
• 2019

자율주행 자동차에서의 보행자 인식 및 사람의 행동 인식과 같은 분야 등에 대한 연구들이 활발하게 진행되고 그에 기반을 둔 기술들이 많이 개발되고 있다. 그리고 대부분의 연구에서는 사람에 대한 경계 박스를 검출한다. 영상에서 사람의 유무 혹은 위치를 판단하는 문제에서는 경계 박스만을 검출하는 것이 효율적일 수 있으나 경계 박스는 행동 인식과 같은 분야에 사용하기에는 많은 정보의 손실이 발생할 수 있다. 본 논문에서는 U-NET 구조의 딥러닝 모델을 사용해 경계 박스로 인한 정보 손실을 줄일 수 있는 보행자 분할 방법을 제안한다. 모델의 학습을 위해 2017 COCO 데이터셋의 사람 카테고리를 사용하였으며 Penn-Fudan 보행자 데이터셋을 이용하여 제안 방법을 테스트하였으며 기존의 방법들과 비교하여 의미 있는 결과를 얻었다.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.24 no.2
• /
• pp.201-208
• /
• 2006

Currently, 1/5,000 Digital Topographic Map covers the whole country with administrative boundary of Gu, while some region has already been constructed with 1/1,000 digital topographic map has legal boundary of Dong. Here, Dong in 1/1,000 scale represents the legal boundary which is used in the address most of time. Therefore, there is no administrative-Dong yet although it is very useful in various fields. In this study, we suggested a method to construct the administrative boundary extending to the level of Dong empirically. In addition, the practical application of the administrative boundary in GIS is discussed. Two methods are applied to construct the administrative boundary to the level of administration-Dong; using the edited cadastral map and the digital topographic map. When the edited cadastral map is used, some problems such as boundary discordance to superordinate administrative area is appeared. On the other hand, using digital topographic map showed simple construction processes and easy connection with other framework data. Therefore, it is recommended that the digital topographic map should be used in the construction of the administrative boundary. It would be useful as framework data in various industries and public operations.

• Journal of KIISE:Software and Applications
• /
• v.37 no.11
• /
• pp.837-844
• /
• 2010

Prediction of the prosodic phrase boundary is one of the most important natural language processing tasks. We propose, for the natural prediction of the Korean prosodic phrase boundary, a statistical approach incorporating efficient learning features. These new features reflect the factors that affect generation of the prosodic phrase boundary better than existing learning features. Notably, moreover, such learning features, extracted according to the hand-crafted prosodic phrase boundary prediction rule, impart higher accuracy. We developed a statistical model for Korean prosodic phrase boundaries based on the proposed new features. The results were 86.63% accuracy for three levels (major break, minor break, no break) and 81.14% accuracy for six levels (major break with falling tone/rising tone, minor break with falling tone/rising tone/middle tone, no break).

• Journal of Korean Society for Geospatial Information Science
• /
• v.20 no.4
• /
• pp.117-125
• /
• 2012

Although many studies have been conducted to extract buildings from airborne lidar data, most of them assume that all the boundaries of a building are straight line segments. This makes it difficult to model building boundaries containing curved segments correctly. This paper aims to model buildings containing curved segments as combination of straight lines and arcs. First, two sets of boundary points are extracted by adaptive convex hull algorithm and local convex hull algorithm with a larger radius. Then, arc segments are determined by average spacing of boundary points and intersection ratio of perpendicular lines. Finally, building boundary is modeled through regularization of least squares line or circle fitting. The experimental results showed that the proposed method can model the curved building boundaries as arc segments correctly by completeness of 69% and correctness of 100%. The approach will be utilized effectively to create automatically digital map that meets the conditions of the Korean digital mapping.

• Journal of KIISE:Software and Applications
• /
• v.28 no.9
• /
• pp.649-661
• /
• 2001

While the conventional edge detection can be considered as the problem of determining the existence of edges at certain locations, the fuzzy edge modeling can be considered as the problem of determining the membership values of edges. Thus, if the location of an edge is unclear, or if the intensity function is different from the ideal edge model, the degree of edgeness at the location is represented as a fuzzy membership value. Using the concept of fuzzy edgeness, an automatic smoothing parameter evaluation and selection method for a conventional edge detector is proposed. This evaluation method uses the fuzzy edge modeling, and can analyze the effect of smoothing parameter to determine an optimal parameter for a given image. By using the selected parameter we can detect least ambiguous edges of a detection method for an image. The effectiveness of the parameter evaluation method is analyzed and demonstrated using a set of synthetic and real images.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.4 no.5
• /
• pp.997-1003
• /
• 2000

When finite length signal is the input of QMF banks which are composed of the M-band orthogonal wavelet filter, the boundaries of reconstructed signal are not reconstructed perfectly. For boundary signals to be reconstructed perfectly, different type filters or methods are applied to them. In this paper, the design method of boundary filters which can be reconstructed the boundary signal perfectly was proposed, in case the dimension of M-band decomposed signal is the same as that of input signal. The boundary filters were designed using the perfect reconstruction condition of paraunitary matrix. In an application of subband coding of still image, the proposed boundary filters achieve better PSNR about 5% in reconstructed image than reflected method at the same bit rate.

• Korean Journal of Cognitive Science
• /
• v.19 no.1
• /
• pp.89-105
• /
• 2008

In this paper, we suggest a rule-based system for the prediction of natural prosodic phrase breaks from Korean texts. For the implementation of the rule-based system, (1) sentence constituents are sub-categorized according to their syntactic functions, (2) syntactic phrases are recognized using the dependency relations among sub-categorized constituents, (3) rules for predicting prosodic phrase breaks are created. In addition, (4) the length of syntactic phrases and sentences, the position of syntactic phrases in a sentence, sense information of contextual words have been considered as to determine the variable prosodic phrase breaks. Based on these rules and features, we obtained the accuracy over 90% in predicting the position of major break and no break which have high correlation with the syntactic structure of the sentence. As for the overall accuracy in predicting the whole prosodic phrase breaks, the suggested system shows Break_Correct of 87.18% and Juncture Correct of 89.27% which is higher than that of other models.

• Journal of KIISE:Software and Applications
• /
• v.33 no.4
• /
• pp.418-426
• /
• 2006

This paper proposes a robust and efficient segmentation method for a cardiac ultrasound image taken from a probe inserted into the heart in surgery. The method consists of three steps: initial boundary extraction, whole boundary modification using confidence competition, and local boundary modification using the rolling spoke method. Firstly, the initial boundary is extracted with threshold regions along the global spokes emitted from the center of an ultrasound probe. Secondly, high confidence boundary edges are detected along the global spokes by competing among initial boundary candidate and new candidates achieved by edge and appearance information. finally, the boundary is modified by rolling local spokes along concave regions that are difficult to extract using the global spokes. The proposed method produces promising segmentation results for the ultrasound cardiac images acquired during surgery.

• Journal of Korean Society for Geospatial Information Science
• /
• v.23 no.2
• /
• pp.11-16
• /
• 2015

In general, problems of delimitating trading area are that it takes much efforts depending on regions and the results are not scientific due to agencies' own rules and criteria. Especially, areas like Hongik University Station, where countless stores are opening and closing simultaneously, the field survey needs even more time and expenses. Despite of its drawbacks, it has been considered that the field survey is the most credible method in delimitating trading area. The purpose of this study, therefore, is reducing the field survey when agencies delimitate trade area by maximizing use of previously built GIS database and utilizing spatial analysis. Publicly notified individual land value and floating population of Gwanak-gu were utilized to delimitate trading areas. For an evaluation method, the study results were compared to other trading area boundaries.

• The KIPS Transactions:PartB
• /
• v.16B no.4
• /
• pp.271-280
• /
• 2009

In this paper, we propose a smoke detection method using region growing method in outdoor video sequences. Our proposed method is composed of three steps; the initial change area detection step, the boundary finding and expanding step, and the smoke classification step. In the first step, we use a background subtraction to detect changed areas in the current input frame against the background image. In difference images of the background subtraction, we calculate a binary image using a threshold value and apply morphology operations to the binary image to remove noises. In the second step, we find boundaries of the changed areas using labeling algorithm and expand the boundaries to their neighbors using the region growing algorithm. In the final step, ellipses of the boundaries are estimated using moments. We classify whether the boundary is smoke by using the temporal information.