Piecewise linear skeletonization using principal curves

被引:145
作者
Kégl, B
Krzyak, A
机构
[1] Univ Montreal, Dept Comp Sci & Operat Res, Montreal, PQ H3C 3J7, Canada
[2] Concordia Univ, Dept Comp Sci, Montreal, PQ H3G 1M8, Canada
基金
加拿大自然科学与工程研究理事会;
关键词
skeletonization; principal curves; feature extraction; image processing;
D O I
10.1109/34.982884
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We propose an algorithm to find plecewise linear skeletons of handwritten characters by using principal curves. The development of the method was inspired by the apparent similarity between the definition of principal curves (smooth curves which pass through the "middle" of a cloud of points) and the medial axis (smooth curves that go equidistantly from the contours of a character image). The central fitting-and-smoothing step of the algorithm is an extension of the polygonal line algorithm [1], [2] which approximates principal curves of data sets by plecewise linear curves. The polygonal line algorithm is extended to find principal graphs and complemented with two steps specific to the task of skeletonization: an initialization method to capture the approximate topology of the character, and a collection of restructuring operations to improve the structural quality of the skeleton produced by the initialization method. An advantage of our approach over existing methods is that we optimize the skeleton graph by minimizing an intuitive and explicit objective function that captures the two competing criteria of smoothing the skeleton and fitting it closely to the pixels of the character image. We tested the algorithm on isolated handwritten digits and images of continuous handwriting. The results indicate that the proposed algorithm finds a smooth medial axis of the great majority of a wide variety of character templates and substantially improves the pixe/wise skeleton obtained by traditional thinning methods.
引用
收藏
页码:59 / 74
页数:16
相关论文
共 29 条
[21]  
Luttrell S P, 1990, IEEE Trans Neural Netw, V1, P229, DOI 10.1109/72.80234
[22]   SKELETONIZATION OF ARABIC CHARACTERS USING CLUSTERING BASED SKELETONIZATION ALGORITHM (CBSA) [J].
MAHMOUD, SA ;
ABUHAIBA, I ;
GREEN, RJ .
PATTERN RECOGNITION, 1991, 24 (05) :453-464
[23]  
NACCACHE NJ, 1984, IEEE T SYSTEMS MAN C, V14
[24]   A THINNING ALGORITHM FOR DISCRETE BINARY IMAGES [J].
PAVLIDIS, T .
COMPUTER GRAPHICS AND IMAGE PROCESSING, 1980, 13 (02) :142-157
[25]  
ROWEIS S, 1998, ADV NEURAL INFORMATI, V10
[26]   Self-organizing maps for the skeletonization of sparse shapes [J].
Singh, R ;
Cherkassky, V ;
Papanikolopoulos, N .
IEEE TRANSACTIONS ON NEURAL NETWORKS, 2000, 11 (01) :241-248
[27]  
SUZUKI S, 1986, P 8 INT C PATT REC, P289
[28]   Probabilistic principal component analysis [J].
Tipping, ME ;
Bishop, CM .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY, 1999, 61 :611-622
[29]   Stochastic jump-diffusion process for computing medial axes in Markov random fields [J].
Zhu, SC .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1999, 21 (11) :1158-1169