Abstract:Stroke extraction and matching are critical for structural interpretation based applications of handwritten Chinese characters, such as Chinese character education and calligraphy analysis. Stroke extraction from offline handwritten Chinese characters is difficult because of the missing of temporal information, the multi-stroke structures and the distortion of handwritten shapes. In this paper, we propose a comprehensive scheme for solving the stroke extraction problem for handwritten Chinese characters. The method consists of three main steps: (1) fully convolutional network (FCN) based skeletonization; (2) query pixel guided stroke extraction; (3) model-based stroke matching. Specifically, based on a recently proposed architecture of FCN, both the stroke skeletons and cross regions are firstly extracted from the character image by the proposed SkeNet and CrossNet, respectively. Stroke extraction is solved by simulating the human perception that once given a certain pixel from non-cross region of a stroke, the whole stroke containing the pixel can be traced. To realize this idea, we formulate stroke extraction as a problem of pairing and connecting skeleton-wise stroke segments which are adjacent to the same cross region, where the pairing consistency between stroke segments is measured using a PathNet [1]. To reduce the ambiguity of stroke extraction, the extracted candidate strokes are matched with a character model consisting of standard strokes by tree search to identify the correct strokes. For verifying the effectiveness of the proposed method, we train and test our models on character images with stroke segmentation annotations generated from the online handwriting datasets CASIA-OLHWDB and ICDAR13-Online, as well as a dataset of Regularly-Written online handwritten characters (RW-OLHWDB). The experimental results demonstrate the effectiveness of the proposed method and provide several benchmarks. Particularly, the precisions of stroke extraction for ICDAR13-Online and RW-OLHWDB are 89.0% and 94.9%, respectively.

A New Stroke-Based Directional Feature Extraction Approach for Handwritten Chinese Character Recognition

HANDWRITTEN CHINESE CHARACTER RECOGNITION WITH DIRECTIONAL DECOMPOSITION CELLULAR FEATURES

Study of Several Handwritten Chinese Character Directional Feature Extraction Approaches

A New Elastic Circle Meshing Feature Extracting Method for Handwriting Chinese Character Recognition

A Feature Extracting Method for Handwritten Chinese Character Recognition Based on Elastic Mesh and Fuzzy Feature of Block

Stroke Extraction of Chinese Character Based on Deep Structure Deformable Image Registration

Off-line Recognition Method for Handwritten Chinese Character Based on Mathematical Morphology

Query Pixel Guided Stroke Extraction with Model-Based Matching for Offline Handwritten Chinese Characters

A New Stroke Extraction Method of Chinese Characters.

An Approach to Natural Stroke Extraction for Off-Line Loosely-Constrained Handwritten Chinese Characters

Analysis and Improvement of Directional Element Feature for Off-Line Handwritten Chinese Character Recognition.

One Stroke Cursive Character Recognition Using Combination of Directional and Positional Features

Handwritten Chinese Character Recognition Based on Sub-stroke

Gabor wavelet based feature extraction for Chinese character recognition

Hidden Markov random field based approach for off-line handwritten Chinese character recognition

A Novel Approach for Stroke Extraction of Off-Line Chinese Handwritten Characters Based on Optimum Paths

Feature Extraction for Handwritten Chinese Character by Weighted Dynamic Mesh Based on Nonlinear Normalization

Feature Extraction for Handwritten Chinese Character by Overlapped Dynamic Meshing and Fuzzy Membership

Triangular Mesh Based Stroke Segmentation for Chinese Calligraphy

Feature Extraction for Online Handwritten Characters Using Delaunay Triangulation

Handwriting Chinese Character Computer Writer Identification Based on Stroke——Computer Writer Identification Based on Motion Feature,First and Last Feature