Abstract
Performing the correct pen-holding gesture plays an important role in handwriting efficiency and quality, especially for early education. In this paper, a detailed design and evaluation of the system, called SmartGe, is presented, which can identify the pen-holding gesture with smartwatch when writing Chinese and English. We firstly analyze the hand movement and propose a novel handwriting detection algorithm to segment each stroke or letter. Then we recognize the pen-holding gesture using deep convolution neural network(DCNN). To improve system performance in Chinese writing, we connect a vertical stroke and a horizontal stroke for pen-holding gesture recognition. SmartGe provides a convenient and natural way to improve users' writing habits, which is a lightweight system, and extensive experiments confirm its effectiveness and robustness.
Highlights
As we know, English and Chinese are two of the most commonly spoken languages in the world
We propose an SmartGe model based on commodity smartwatches, which can detect pen-holding gestures on the paper
Our proposed SmartGe system proves that people can leverage the inertial sensors built-in smartwatch to recognize pen-holding gesture, which is robust to lighting conditions and can avoid interference of wireless signals compared with existing works
Summary
English and Chinese are two of the most commonly spoken languages in the world. We present the design, implementation and evaluation of SmartGe, which can detect pen-holding gesture leveraging built-in sensors of smartwatches when writing. Our proposed SmartGe system proves that people can leverage the inertial sensors built-in smartwatch to recognize pen-holding gesture, which is robust to lighting conditions and can avoid interference of wireless signals compared with existing works. OVERVIEW we will outline the system design of SmartGe. The system is designed for recognizing penholding gesture in both English and Chinese writing only relying on the smartwatch, which can distinguish 9 different types of pen-holding gestures, including one correct and eight incorrect [14]. To build a more accurate classification model for recognizing handwriting gestures, we develop a handwriting detection algorithm to identify the event of writing by using the signals of strokes or letters as input to avoid detecting incomplete or redundant signals. No matter what characters or words we write, the pen-holding gesture can be detected by us normally
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have