본문 바로가기

Linking soft computing to art: introduction of efficient k-continuous line drawing

문서 유형
연구출판물 · 학술지 논문
등재
SCIEScopus
저널 순위
JCR Q1Scopus Q1
저널
APPLIED SOFT COMPUTING68 · pp.932 · 2018
피인용
Web of Science 0Scopus 1
FWCI (Scopus)
0
JCR IF
JCR 2017 3.907
ISSN
1568-4946

저자

  1. 주염걸제1저자Department of Industrial Engineering, Pusan National University, Busan, South Korea
  2. 이규민교신저자Department of Industrial Engineering, Pusan National University, Busan, South Korea

초록

Continuous line drawing (CLD) is a technique used in the field of art, in which the pen does not leave the paper until the sketch is completed. In this study, a novel technique, k-continuous line drawing (kCLD), is proposed; this technique enables a visual image to comprise k closed non-intersecting lines. k-CLD involves the following challenges: 1) partitioning the target image into k regions, 2) stippling each region without distorting the target image, and 3) connecting the stippled dots in each region using a single closed, non-intersecting line. This study identifies and implements efficient algorithms to produce high-quality k-CLDs. Further, an improvement to the graph-based image segmentation algorithm has been proposed using the Minkowski distance to evaluate dissimilarity difference and demonstrated its effectiveness to partition the target image into k regions. Next, well-spaced stippled dots were generated in each region using a weighted Voronoi diagram. Finally, the stippled dots were connected by a single non-intersecting line, obtained by solving a traveling salesman problem (TSP) in each region. The meta heuristic to solve the TSP was an ant colony system algorithm. The proposed methodologies were tested on a wide variety of images to demonstrate their effectiveness and efficiency. (C) 2017 Elsevier B.V. All rights reserved.

키워드

k-Continuous line drawingImage segmentationTraveling salesman problemWeighted voronoi diagramAnt colony system algorithmMinkowski distanceTRAVELING SALESMAN PROBLEMIMAGE SEGMENTATIONALGORITHM

식별자

Scopus EID
2-s2.0-85039069774
WoS UID
000433155300070