——商業(yè)化的產(chǎn)品案例，引自DTing-虎嗅網(wǎng)
產(chǎn)品簡介
DTing是一款感知交互設(shè)備，主要應(yīng)用于手勢識別，它依靠的是表面肌肉電信號和慣性原件技術(shù)，能通過裹在人體前臂上的傳感器來監(jiān)測手指牽引的相應(yīng)肌肉群，并依靠核心算法翻譯、轉(zhuǎn)化成相應(yīng)的信號輸出。
DTing將肌肉電傳感技術(shù)和人工智能技術(shù)相結(jié)合，建立了生物信號在醫(yī)學和健康領(lǐng)域以外的創(chuàng)新型應(yīng)用----人機交互。我們使用前臂肌肉群的表面肌肉電信息來進行手勢的識別，既避免了現(xiàn)有機器視覺識別方式對于使用環(huán)境和場景的限制，又可以在使用的過程中避免占用雙手。使得DTing可以實現(xiàn)手指手勢、手腕動作和前臂動作的綜合識別，在3D空間中進行。
基于高分辨率的電阻抗成像提高手勢識別
Abstuct
Electrical Impedance Tomography (EIT，電阻抗成像) was recently employed in the HCI domain to detect hand gestures using an instrumented smartwatch.
This prior work demonstrated great promise for non-invasive, high accuracy recognition of gestures for interactive control（手勢識別的交互）. We introduce a new system that offers improved sampling speed and resolution. In turn, this enables superior interior reconstruction and gesture recognition.
More importantly, we use our new system as a vehicle for experimentation – we compare two EIT sensing methods and three different electrode resolutions. Results from in-depth empirical evaluations and a user study shed light on the future feasibility of EIT for sensing human input.
總結(jié)：EIT的成像原理類同CT，本文基于不同動作肌肉群的成像不同做手勢的識別。結(jié)合了CNN+圖像-》分類問題的套路，正確率也不錯。
Zhang Y, Xiao R, Harrison C. Advancing Hand Gesture Recognition with High Resolution Electrical Impedance Tomography[C]//Proceedings of the 29th Annual Symposium on User Interface Software and Technology. ACM, 2016: 843-850.
轉(zhuǎn)載鏈接：https://zhuanlan.zhihu.com/p/26010511
蘇公網(wǎng)安備32011202000907