Abstract:Recently, the hybrid convolutional neural network hidden Markov model (CNN-HMM) has been introduced for offline handwritten Chinese text recognition (HCTR) and has achieved state-of-the-art performance. However, modeling each of the large vocabulary of Chinese characters with a uniform and fixed number of hidden states requires high memory and computational costs and makes the tens of thousands of HMM state classes confusing. Another key issue of CNN-HMM for HCTR is the diversified writing style, which leads to model strain and a significant performance decline for specific writers. To address these issues, we propose a writer-aware CNN based on parsimonious HMM (WCNN-PHMM). First, PHMM is designed using a data-driven state-tying algorithm to greatly reduce the total number of HMM states, which not only yields a compact CNN by state sharing of the same or similar radicals among different Chinese characters but also improves the recognition accuracy due to the more accurate modeling of tied states and the lower confusion among them. Second, WCNN integrates each convolutional layer with one adaptive layer fed by a writer-dependent vector, namely, the writer code, to extract the irrelevant variability in writer information to improve recognition performance. The parameters of writer-adaptive layers are jointly optimized with other network parameters in the training stage, while a multiple-pass decoding strategy is adopted to learn the writer code and generate recognition results. Validated on the ICDAR 2013 competition of CASIA-HWDB database, the more compact WCNN-PHMM of a 7360-class vocabulary can achieve a relative character error rate (CER) reduction of 16.6% over the conventional CNN-HMM without considering language modeling. By adopting a powerful hybrid language model (N-gram language model and recurrent neural network language model), the CER of WCNN-PHMM is reduced to 3.17%.

From Textline to Paragraph: A Promising Practice for Chinese Text Recognition

Recognition of Handwritten Chinese Text by Segmentation: A Segment-annotation-free Approach

A New Hybrid-Parameter Recurrent Neural Network for Online Handwritten Chinese Character Recognition

An approach for handwritten Chinese text recognition unifying character segmentation and recognition

A Handwritten Chinese Text Recognizer Applying Multi-level Multimodal Fusion Network

Sentence-level Online Handwritten Chinese Character Recognition

Fully Convolutional Recurrent Network for Handwritten Chinese Text Recognition

Writer-Aware CNN for Parsimonious HMM-Based Offline Handwritten Chinese Text Recognition

PageNet: Towards End-to-End Weakly Supervised Page-Level Handwritten Chinese Text Recognition

Offline Handwritten Chinese Text Recognition with Convolutional Neural Networks

Learning Spatial-Semantic Context with Fully Convolutional Recurrent Network for Online Handwritten Chinese Text Recognition

Handwritten Chinese Text Recognition by Integrating Multiple Contexts

SegCTC: Offline Handwritten Chinese Text Recognition Via Better Fusion Between Explicit and Implicit Segmentation

Handwriting Text-line Detection and Recognition in Answer Sheet Composition with Few Labeled Data

Deep LSTM Networks for Online Chinese Handwriting Recognition

Towards Fast, Accurate and Compact Online Handwritten Chinese Text Recognition

A Transfer Learning Method for Handwritten Chinese Character Recognition

A Fast and Accurate Fully Convolutional Network for End-to-End Handwritten Chinese Text Segmentation and Recognition

Transcript Mapping for Handwritten Chinese Documents by Integrating Character Recognition Model and Geometric Context

A Hybrid Post-Processing System for Handwritten Chinese Character Recognition

Towards End-to-end Handwritten Document Recognition