研究目的
Investigating the use of gyroscopes and accelerometers in smartphones for reconstructing the motion path for pen-like interaction to enable text entry.
研究成果
GyroPen is a promising approach for handwriting with smartphones, enabling text entry without the need for a stylus or touchscreen. The method reconstructs writing paths accurately enough for recognition by an off-the-shelf handwriting recognition engine, with users able to write at speeds of 3–4 s per word and a character error rate of 18% after some practice.
研究不足
The method is sensitive to noise in sensor data, and the sensors in current phones may be too noisy for accurate estimation of a writing path. The system does not handle 'pen up' movements explicitly, which can lead to connected strokes in the reconstructed path.
1：Experimental Design and Method Selection:
The study uses gyroscopes and accelerometers in smartphones to reconstruct the trajectory of the phone’s corner touching a writing surface. The method involves direct use of angular trajectory for reconstruction.
2：Sample Selection and Data Sources:
The experiments involved novice and experienced users writing words and drawings, with data collected from the phone’s sensors.
3：List of Experimental Equipment and Materials:
Samsung Galaxy Nexus and Samsung Nexus S smartphones with Invensense MPU-3050 motion sensing unit and EM-Tech EME1511AFRC module.
4：Experimental Procedures and Operational Workflow:
Users wrote words and drawings with the phone held like a pen, and the system reconstructed the writing path from sensor data.
5：Data Analysis Methods:
The reconstruction accuracy was evaluated by comparing the reconstructed paths with ground-truth paths from a graphic tablet, and the character error rate (CER) was measured for text entry.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
