Oguz Yetkin, Karmina Calderon, P. Krishna Moorthy, Thao Thu Nguyen, Jennifer Tran, Taylor Terry, Anthony Vigil, Anne Alsup, Aaron Tekleab, Danielle Sancillo, Nosisa Ncube, J. Baptist
{"title":"A Lightweight Wearable American Sign Language Translation Device","authors":"Oguz Yetkin, Karmina Calderon, P. Krishna Moorthy, Thao Thu Nguyen, Jennifer Tran, Taylor Terry, Anthony Vigil, Anne Alsup, Aaron Tekleab, Danielle Sancillo, Nosisa Ncube, J. Baptist","doi":"10.1115/dmd2022-1053","DOIUrl":null,"url":null,"abstract":"\n A lightweight wearable American Sign Language translation device is presented consisting of optically communicating rings worn on each finger and wireless devices worn on each fingernail. The device is similar in principle to glove based sign language translation devices except it does not require the user to wear an entire glove. Each ring is constructed from optically clear resin containing three infrared transponders, controlled by a microcontroller unit worn on the wrist. The fingernail units are also cast in clear resin and contain circuitry to receive and respond to infrared light, and house three small SR60 watch batteries. The system uses a microcontroller to flash all LEDs in succession and create matrix sensor readings from each LED, resulting in a 15x15 matrix corresponding to a signed gesture, recognized by an Artificial Neural Network. This system is expected to be able to recognize 24 of the 26 letters of the ASL alphabet and be extended to recognize arbitrary ASL signs with the integration of an accelerometer. The system currently recognizes 9 letters (A, B, D, F, K, N, O, T, X) with intention to build on this novel method of gesture recognition to construct a full-fledged ASL translator. The system is currently useable as a low-encumbrance gesture-based input system for arbitrary device control.","PeriodicalId":236105,"journal":{"name":"2022 Design of Medical Devices Conference","volume":"96 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 Design of Medical Devices Conference","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1115/dmd2022-1053","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
A lightweight wearable American Sign Language translation device is presented consisting of optically communicating rings worn on each finger and wireless devices worn on each fingernail. The device is similar in principle to glove based sign language translation devices except it does not require the user to wear an entire glove. Each ring is constructed from optically clear resin containing three infrared transponders, controlled by a microcontroller unit worn on the wrist. The fingernail units are also cast in clear resin and contain circuitry to receive and respond to infrared light, and house three small SR60 watch batteries. The system uses a microcontroller to flash all LEDs in succession and create matrix sensor readings from each LED, resulting in a 15x15 matrix corresponding to a signed gesture, recognized by an Artificial Neural Network. This system is expected to be able to recognize 24 of the 26 letters of the ASL alphabet and be extended to recognize arbitrary ASL signs with the integration of an accelerometer. The system currently recognizes 9 letters (A, B, D, F, K, N, O, T, X) with intention to build on this novel method of gesture recognition to construct a full-fledged ASL translator. The system is currently useable as a low-encumbrance gesture-based input system for arbitrary device control.
介绍了一种轻型可穿戴美国手语翻译设备,该设备由每个手指上佩戴的光通信环和每个指甲上佩戴的无线设备组成。该设备在原理上类似于基于手套的手语翻译设备,除了它不需要用户戴上整个手套。每个环由光学透明树脂制成,包含三个红外转发器,由佩戴在手腕上的微控制器单元控制。指甲单元也用透明树脂铸造,包含接收和响应红外光的电路,并容纳三个小型SR60手表电池。该系统使用微控制器连续闪烁所有LED,并从每个LED创建矩阵传感器读数,从而产生与签名手势对应的15x15矩阵,由人工神经网络识别。该系统预计能够识别26个美国手语字母中的24个,并通过集成加速度计扩展到识别任意美国手语符号。该系统目前可以识别9个字母(A, B, D, F, K, N, O, T, X),并打算在这种新颖的手势识别方法的基础上构建一个完整的美国手语翻译。该系统目前可作为一种低障碍的基于手势的输入系统,用于任意设备控制。