輔助記憶認知障礙者之穿戴裝置的設計

Design of the wearable device to assist the memory of people with cognitive impairment

林聖梅(Sheng-Mei Lin)

指導教授 ： 陳建霖

義守大學/理工學院/機械與自動化工程學系/碩士(2019年)

摘要

本研究利用機械學習技術，輔助人臉記憶以協助使用者喚起記憶，避免尷尬之心理不平衡。本研究是設計製作一個整合型的穿載式裝置，包含基座、資料處理器、攝影機與耳機；此裝置以帽子型態呈現，並將影音數據傳至處理器中進行人臉辨識及運算。本研究先運用電腦輔助設計技術(CAD)設計裝置系統，並運用電腦輔助工程分析(CAE)求解基座的應力、應變與位移量，藉以改良裝置之設計，然後運用3D列印技術列印本裝置。本研究的系統運作如後：首先運用OpenCV的Haar分類器偵測人臉位置，然後利用LBPH人臉識別器進行交談者的身分識別，最後處理器啟動語音檔回應此人臉的資料，達到即時告知記憶認知障礙者所面對交談者的身分。本研究運用CAD、CAE及3D 列印技術來設計穿戴裝置，因此有客製化與即時製作的特點，所以適用於醫療輔具的開發。本輔助記憶裝置經過實驗室內的測試，證實易於使用者穿戴；並可依使用者的需求，建立所需的辨識者資料庫，是未來可用於記憶認知障礙者的一種輔具。

關鍵字

3D列印 ； 人臉辨識

並列摘要

This study uses the machine learning technique to assist face recognition to help users evoke memories and avoid mental imbalances. This study is to design and manufacture an integrated wearable device, including the support, data processor, camera and earphone. The device is presented in a hat style, and it can transmit the audio and video data to the processor for face recognition and calculations. This study first uses computer-aided design technology (CAD) to design the device system, and uses computer-aided engineering analysis (CAE) to carry out the stress, strain and displacement of the support. In addition to improve the device design, 3D printing technology are employed to print the device. The operation of the present system is as follows: Firstly, the Haar classifier of OpenCV is used to detect the face position, and then the LBPH face recognizer is used to identify the identity of the talker. Finally, the voice response will be activated to provide the profile of the talker to the wearer. Therefore, the instant notification is realized by the people with memory cognitive impairment. The present study uses CAD, CAE and 3D printing technologies to design this wearable device, so it has the characteristics of customization and instant manufacturing. We believe it is suitable for the development of medical aids. This memory-aided device has been tested in the laboratory to prove that it is easy for the user to wear. Moreover, it can establish a personal identification database according to the users’ needs, and could be an assistive device for the people with memory cognitive impairment in the future.

並列關鍵字

CAD ； CAE ； 3D printing ； Face recognition.

參考文獻

[1] 陳杰翰，莊坤達，人工智慧的非技術挑戰，科學發展月刊，2019。

[2] 詹峻陽，人工智慧商業化的革命時刻，數位時代月刊NO.270，2016。來自此https://www.bnext.com.tw/article/41534/3-key-techniques-of-ai

[3] C. García-Saura and J. González-Gómez, “Low cost educational platform for robotics, using open-source 3d printers and open-source hardware,” ICERI2012 proceedings, 2012.

[4] J. Gwamuri, J. E. Poliskey, and J. M. Pearce, “Open Source 3-D Printers: An Appropriate Technology For Developing Communities,” 2017.

[5] Moorhead, Joanna. Raspberry Pi device will 'reboot computing in schools'. The Guardian, 2012.