Most handwritten Chinese character recognition systems suffer from the vari
ations in geometrical features for different writing styles. The stroke str
uctures of different styles have proved to be more consistent than geometri
cal features. In an on-line recognition system, the stroke structure can be
obtained according to the sequences of writing via a pen-based input devic
e such as a tablet. But in an off-line recognition system, the input charac
ters are scanned optically and saved as raster images, so the stroke struct
ure information is not available. In this paper, we propose a method to ext
ract strokes from an off-line handwritten Chinese character. We have develo
ped four new techniques: 1) a new thinning algorithm based on Euclidean dis
tance transformation and gradient oriented tracing, 2) a new line approxima
tion method based on curvature segmentation, 3) artifact removal strategies
based on geometrical analysis, and 4) stroke segmentation rules based on s
plitting, merging and directional analysis. Using these techniques, we can
extract and trace the strokes in an off-line handwritten Chinese character
accurately and efficiently.