Name
Abstract | ||
|---|---|---|
American Sign Language (ASL) is widely used among hearing impaired individuals in English-speaking countries. Various technologies have been developed to perform ASL recognition, including optical signal sensing, electrical signal sensing, and mechanical signal sensing. However, wearable devices using those methods have bulky and complex sensing modules that lead to long-term discomfort as well as poor accuracy. In this paper, we present an epidermal-iontronic sensing (EIS)-based wearable device that wears on finger joints for 35 fingerspelling ASL recognitions (i.e., 26 alphabets from A to Z and 9 digits from one to nine). Compared to current on-market devices, such design is lighter, comfortable to wear and has better appearance according to user comments. When bending the finger, a physical contact forms between the ionic material and the epidermis of skin, leading to an electric double layer (EDL) established at the interface. Therefore, a significant capacitive change can be achieved with various finger gestures. By using Nafion as the ionic sensing material, we developed a sensing device to provide excellent flexibility and optical transparency. We used machine learning methods, such as neural networks to track and perform ASL recognition using the signals obtained from the designed device. The algorithm achieved a within-user accuracy of 99.6% and a cross-user accuracy of 76.1% when adapted the model to different users. This wearable device is low-cost and has broad potential to be integrated in future application of human-machine interactions (HMI), smart home controls, and nonverbal communications.
|
Year | DOI | Venue |
|---|---|---|
2018 | 10.1145/3287080 | Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies |
Keywords | DocType | Volume |
American Sign Language,Gesture Recognition,Input Techniques,Iontronic Capacitive Sensing,User Input,Wearable Device | Journal | 2 |
Issue | ISSN | Citations |
4 | 2474-9567 | 1 |
PageRank | References | Authors |
0.36 | 0 | 6 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Zijie Zhu | 1 | 3 | 0.77 |
| Xuewei Wang | 2 | 72 | 9.72 |
| Aakaash Kapoor | 3 | 1 | 0.36 |
| Zhichao Zhang | 4 | 6 | 4.88 |
| Tingrui Pan | 5 | 10 | 7.22 |
| Zhou Yu | 6 | 278 | 39.88 |