Abstract:A novel framework (Stroke‐Seg) specifically designed to improve the stroke segmentation of Chinese characters is proposed by tackling complex stroke characteristics, and the framework can be easily integrated with most networks. The proposed framework, utilizing a dedicated 'Stroke Class' mechanism, prior knowledge vector, multi‐label output strategy, and BDLoss optimization, significantly enhances Chinese character stroke segmentation accuracy and generalization ability across various architectures while reducing computational cost. Chinese stroke segmentation is a crucial and challenging task for various downstream applications such as font generation, aesthetic evaluation etc. Conventional semantic segmentation techniques typically face difficulties in accurately segmenting strokes, as intersection regions in Chinese characters can belong to multiple strokes simultaneously, and these approaches often lack a holistic understanding of character composition. This paper proposes a character stroke segmentation framework named Stroke‐Seg that integrates with various semantic segmentation architectures, demonstrating adaptability to different backbone networks to tackle the above tasks. A multi‐label output strategy is proposed to effectively classify strokes in intersection areas, overcoming the limitations of traditional semantic segmentation approaches. Additionally, a prior knowledge vector is incorporated into the input layer to provide character‐specific information on stroke composition, enhancing the ability of the framework to precisely identify and segment strokes. The effectiveness of the proposed framework is demonstrated through evaluations of a comprehensive dataset (brush calligraphy stroke segmentation dataset). Evaluations show that the proposed framework significantly improves the capability of semantic segmentation networks, achieving a remarkable improvement of up to 19.9% in the true stroke rate, compared to traditional stroke segmentation techniques. The Stroke‐Seg framework integrated with TransUNet demonstrates its high performance with an impressive 98.2% true stroke rate. Furthermore, the proposed framework combined with FCN still achieves good performance while consuming the least amount of computational resources and memory, demonstrating its potential for lightweight design.

Segmentation of Numeral Strings Using Stroke Grouping

Research and Implementation of Structural Features of Characters Based on a Soft K-Segments Algorithm for Principal Curves

A Recognition Based System for Segmentation of Touching Handwritten Numeral Strings

Segmentation of Online Sketching Using Geometric Feature

Constraint reconstruction for sectional curve of extruded and rotational surface using improved methods

Off-Line Handwritten Digit Recognition Based on Principal Curves

An Online Handwritten Numerals Segmentation Algorithm Based on Spectral Clustering.

Principal curves algorithm based on subdivision

Snowvision: Segmenting, Identifying, and Discovering Stamped Curve Patterns from Fragments of Pottery

A Novel Approach Of Segmenting Touching And Kerned Characters

Segmentation of Crossing Planar Curves

Statistical Modeling and Learning for Recognition-Based Handwritten Numeral String Segmentation.

A Robust Algorithm of Principal Curve Detection

Triangular Mesh Based Stroke Segmentation for Chinese Calligraphy

Stroke‐Seg: A deep learning‐based framework for chinese stroke segmentation

The Approach of Grouping Straight Line Segments Based on Perceptual Organization

Threshold Segmentation and Length Measurement Algorithms for Irregular Curves in Complex Backgrounds

Segmentation of Online Freehand Stroke Using Geometrical Feature

A New Handwritten Character Segmentation Method Based on Nonlinear Clustering

A 3D Curve Grouping Methods and Its Application

A top-down algorithm for finding principal curves from complex patterns data