Hiroyuki Ikezoe
poinc****@ikezo*****
2007年 4月 12日 (木) 12:14:13 JST
Hi, 2007-04-12 (木) の 08:37 +0800 に Huang Peng さんは書きました: > On Wed, 2007-04-11 at 17:33 +0800, Hu Zheng wrote: > > As build the stroke info manually is very troublesome, I want to discuss > > about whether it is possible to generate stroke data from ttf font file. > > I will research this idea tomorrow, by read ttf spec and freetype source > > code. > Hi Hu, > IMHO, ttf is vector font, but it only has data of charaters' outlines, > it does not split a character to several strokes, and does not have > strokes related data. I think it need complex algorithm to generate > strokes from it. :( I agree with you completely. TOMOE's recognition engine is an online handwriting recognition. It needs each stroke data. Moreover, when TOMOE has a HMM engine or something for recognition, all strokes data should be resampled even if existing TOMOE data can be used. There is no easy way to achieve nice work. :-) -- Hiroyuki Ikezoe
