Incremental learning of LDA model for Chinese writer adaptation

A new writer adaption method based on incremental linear discriminant analysis (ILDA) is presented in this paper. We first provide a more general solution for ILDA and then present a Weighted ILDA (WILDA) approach. Based on ILDA or WILDA, the writer adaptation is performed by updating the LDA transformation matrix and the classifier prototypes in the discriminative feature space. Experimental results show that both ILDA and WILDA are very effective to improve the recognition accuracy for writer adaptation, and WILDA outperforms ILDA. The proposed WILDA based writer adaptation method can reduce as much as 47.88% error rate on the writer-dependent dataset while it only has as less as 0.85% accuracy loss on the writer-independent dataset. It indicates that writer adaption using WILDA can significantly increase the recognition accuracy for the particular writer while having limited impact on the accuracy for general writers.