A reduced classifier ensemble approach to human gesture classification for robotic Chinese handwriting

Abstract

The paper presents an approach to applying a classifier ensemble to identify human body gestures, so as to control a robot to write Chinese characters. Robotic handwriting ability requires complicated robotic control algorithms. In particular, the Chinese handwriting needs to consider the relative positions of a character's strokes. This approach derives the font information from human gestures by using a motion sensing input device. Five elementary strokes are used to form Chinese characters, and each elementary stroke is assigned to a type of human gestures. Then, a classifier ensemble is applied to identify each gesture so as to recognize the characters that gestured by the human demonstrator. The classier ensemble's size is reduced by feature selection techniques and harmony search algorithm, thereby achieving higher accuracy and smaller ensemble size. The inverse kinematics algorithm converts each stroke's trajectory to the robot's motor values that are executed by a robotic arm to draw the entire character. Experimental analysis shows that the proposed approach can allow a human to naturally and conveniently control the robot in order to write many Chinese characters.

Use of Automatic Chinese Character Decomposition and Human Gestures for Chinese Calligraphy Robots

“…The stroke learning module is responsible for the learning of the stroke writing skills for the robot. This module is developed based on previous work regarding learning robotic writing ability though human gesture analysis [22]. In order to perform Chinese character writing, five emblematic command gestures are chosen to represent five primitive/elementary Chinese strokes.…”

“…almost all existing research applied predefined font databases, which were created by computational programming [12], [13] or by follow-up mechanisms [14], [22]; on the following four important features: 1) font database implementation methods, 2) ability to write new characters, 3) ability to write various font styles, and 4) human-robot interactions.…”

Integration of an actor-critic model and generative adversarial networks for a Chinese calligraphy robot

“…In response to this problem, several researchers use human aesthetic feedback as the evaluation criterion [12,13]. For example, Chao et al developed a human-robot interaction model to evaluate writing results [14,15]. Other researchers manually designed evaluation methods to evaluate the quality of calligraphy [16,17,18].…”

GANCCRobot: Generative adversarial nets based chinese calligraphy robot

“…The first challenge is often handled by the employment of large stroke datasets to train robot control systems [14]. These methods indeed improves the style diversity of the generated strokes to some extent, but the challenge remains as the styles are still tied with the training samples [15,16]. Notice that calligraphers break through the limitation of learning samples and create new writing styles through human creativity, and human-computer interaction methods have been widely used in robotic control [17,18].…”