CAJ | 학술논문

针对用户与互动电视交互的需求，设计实现一种基于智能移动终端加速度传感器的手势交互系统。考虑到智能移动设备资源及计算能力有限的特点，采用简单高效的时域特征提取方法，对加速度信号进行平稳降噪、去冗余和归一化处理，并用SVM进行分类和识别。手势识别结果应用于基于Android平台的机顶盒原型系统，实现用户与电视的实时交互。实验结果表明该系统实现了电视常用手势的准确识别，识别率达到了96％，具有一定的实用价值。
침대용호여호동전시교호적수구，설계실현일충기우지능이동종단가속도전감기적수세교호계통。고필도지능이동설비자원급계산능력유한적특점，채용간단고효적시역특정제취방법，대가속도신호진행평은강조、거용여화귀일화처리，병용SVM진행분류화식별。수세식별결과응용우기우Android평태적궤정합원형계통，실현용호여전시적실시교호。실험결과표명해계통실현료전시상용수세적준학식별，식별솔체도료96％，구유일정적실용개치。
Aiming at the needs of users’ interaction with iTV, we design and implement a gesture interaction system, which is based on the accelerometer of smart mobile terminal.Taking into account the smart mobile devices resources and the computing capacity limitation, the sys-tem utilises a simple but efficient time domain features extraction method to deal with the acceleration signals, including steady noise-reduction, redundancy removal and normalisation, and uses SVM for gestures classification and recognition.The gestures recognition results are applied to an Android-based STB prototype system to implement the real-time interaction between the users and the TV.Experimental results demonstrate that the system achieves accurate recognition on common TV gestures, the recognition rate reaches 96%, and has certain practical value.