Calligraphy imitation (CI) from a handful of target handwriting samples is
such a challenging task that most of the existing writing style analysis or
handwriting generation methods do not exhibit satisfactory performance. In this paper, we propose a novel multi-module framework to address the problem of CI. Firstly, we utilized a deep convolution neural network (CNN) to extract personalized calligraphical features. Then we built a calligraphyclustering attention module and a mata-style matrix (msM) to compute an
embedding of calligraphy. The structure of conditional gated recurrent unit (cGRU) is then improved to predict the probabilistic density of pen tip movement displacement by dual condition inputs. Finally, we generated personalized handwriting stroke sequences through iterative sampling with Gaussian mixture model (GMM). Experiments on public online handwriting databases verify that the proposed method could achieve satisfactory performance; the generated samples achieved high similarities with original handwriting examples.