未來展望

以個人電腦和 ZedBoard 嵌入式開發板執行眼動儀系統程式，比較兩 者之間的系統執行時間。系統執行時間比較分成兩個部份，分別為九點校 正時間與計算凝視點時間。九點校正時間是螢幕眼動儀進行校正時，使用 者凝視電腦螢幕上預設已知凝視點，即螢幕上的九點座標位置，當使用者 凝視已知座標點時，需要使用滑鼠點擊座標點之按鈕，此時的執行時間個 人電腦與嵌入式平台的計算時間。如表 5.1 所示，為兩平台的九點校正計 算時間比較表，從表中可得知個人電腦的執行時間比嵌入式平台快，平均 相差 10.26 倍。

表 5.1 個人電腦與嵌入式平台之九點校正執行時間比較表

螢幕座標位置 個人電腦 嵌入式平台

calibration_1 0.288 ms 3.051 ms calibration_2 0.335 ms 3.346 ms calibration_3 0.387 ms 3.396 ms calibration_4 0.342 ms 3.889 ms calibration_5 0.355 ms 3.281 ms calibration_6 0.307 ms 3.806 ms calibration_7 0.357 ms 3.525 ms calibration_8 0.346 ms 3.617 ms calibration_9 0.358 ms 3.632 ms average 0.341 ms 3.500 ms

由於進行九點校正的過程，使用者凝視螢幕上的預設座標點，之後點 擊滑鼠左鍵，取瞳攝影機記錄此時瞳孔中心的位置，並且把預設的座標點 存入螢幕眼動儀程式中。因此執行時間為點擊預設座標點，紀錄資料的時 間，不包含使用者凝視座標點的時間，因不同的使用者的凝視時間完全不 相同，故把這項因素排除在研究之外。

九點校正的執行時間，是根據取瞳攝影機所拍攝的瞳孔中心影像，與

從實驗設備進行分析，個人電腦的中央處理器的規格為 Intel Core i5-5200 CPU @ 2.20GHz、Quad-Core；而嵌入式平台的中央處理器規格為 ARM Cortex-A9 @ 667MHz、Dual-Core，兩者的處理器規格相差較大，從 處理器的頻率來看，兩者頻率相差近 4 倍，核心數則是處理器個人電腦為 四核心，嵌入式平台為雙核心，若單純比較 CPU 之硬體，可明顯知道兩者 差異性，因此程式之執行時間有落差。

系統執行時間實驗為比較兩平台執行九點校正與計算凝視點的時間，

在九點校正的比較實驗中，點擊螢幕上已知的九點位置，紀錄眼睛之瞳孔 座標數值所執行的時間，在個人電腦所執行的時間為 0.341 毫秒，嵌入式 平台所執行的平均時間為 3.50 毫秒，比較兩者的執行時間，結果為個人電 腦比嵌入式平台快 10.26 倍。在計算凝視點的時間上，取得眼睛瞳孔的數 據與預設螢幕座標點後，計算出二次多項式的各項參數，根據這些參數可 計算出凝視點的位置。個人電腦所執行的時間為 0.6 毫秒；嵌入式平台所 執行的時間為 59 毫秒。比較兩者的執行時間，結果為個人電腦比嵌入式 平台快 98.33 倍。因此針對提升此嵌入式系統之眼動儀及嵌入式開發板的 應用，在未來研究方向建議如下列三點陳述：

1. 使用 ZedBoard 的可程式化邏輯，設計眼動儀的硬體模組。

由於影像資訊取得需要使用較大的計算量，可透過硬體加速 的方式，把此工作分配給專門的硬體處理，可以減輕中央處理器 的工作量，以空間換取時間的硬體設計，可使用 ZedBoard 的可程 式化邏輯部分，即為 FPGA 上的硬體設計，增加相同的硬體資源，

如邏輯面積、空間，進行平行處理，實現處理效能成長。

2. 完成眼動儀之頭動補償的設計。

目前此系統仍需固定頭部才能減少系統誤差，因眼動儀的校 正與操作，眼睛和電腦螢幕的距離必須固定，否則會產生凝視誤

差。若此嵌入式系統環境加上頭動補償功能，可能會降低處理速 度，未來可考量是否循找到適合的演算法進行修改。

3. 減少個人操作的流程。

目前的使用流程先找到瞳孔中心，再進行九點校正，才能使 用此螢幕眼動儀系統。未來可減少系統自行操作的部分，如自動 尋找瞳孔中心，簡化個人操作步驟，讓系統更人性化。

參考文獻

[1] Andrew T. Duchowski, “A Breadth First Survey of Eye Tracking Applications,” Behavior Research Methods, Instruments, & Computers, vol. 34, no. 4, pp. 455-470, Nov. 2002.

[2] Karel Fliegel, “Eyetracking based approach to objective image quality assessment,” International Carnahan Conference on Security Technology, Prague, Oct. 2008, pp.371-376.

[3] Areej Al-Wabil, Ebtisam Alabdulqader, Latifa Al-Abdulkarim, and Nora Al-Twairesh, “Measuring the User Experience of Digital Books with Children: An Eyetracking Study of Interaction with Digital Libraries,”

Internet Technology and Secured Transactions, London, Nov. 2010, pp.1-7.

[4] W Lemahieu, B Wyns, “Low cost eye tracking for human machine interfacing,” Journal of Eyetracking, Visual Cognition and Emotion, vol. 1, no. 1, pp. 1-12, Oct. 2010.

[5] Kuryati Kipli, Theo Arvanitis, Neil Cooke, and Lisa Harris, “An eyetracking study of estimation accuracy: Examining cerebellar tumours from Magnetic resonance spectroscopy graphs,” International Symposium on Information Technology, Kuala Lumpur, Aug. 2008, pp.1-5.

[6] William Garrard, “An examination of eyetracking artifacts related to website design for individuals with cognitive disability,” International Conference on Systems, Man, and Cybernetics (SMC), San Diego, Oct.

2014, pp.3197-3202.

[7] Anirban Chowdhury, Sougata Karmakar, Swathi Matta Reddy, Sanjog J., Subrata Ghosh, and Debkumar Chakrabarti, “Visual Attention Analysis on Mutated Brand Name using Eye-Tracking: A Case Study,” World Academy of Science, Engineering and Technology, India, 2012, pp.1132-1135.

[8] Mihajlov Martin, Trpkova Marija, and Arsenovski Sime “Eye tracking recognition-based graphical authentication,” Application of Information and Communication Technologies, Baku, Oct. 2013, pp.1-5.

[9] 江宗憲，“低成本高速眼動儀之建構”，國立臺灣師範大學應用電子研 究所，碩士論文，2013。

[10] 曾士誠，“頭戴式眼動儀之頭動補償探討”，國立臺灣師範大學電機工 程學系，碩士論文，2015。

[11] Shahram Eivazi, Roman Bednarik, Ville Leinonen, Mikael von und zu Fraunberg, and Juha E. Jaaskelainen, “Embedding an Eye Tracker Into a Surgical Microscope: Requirements, Design, and Implementation,” IEEE Sensors Journal, vol. 16, no. 7, pp. 2070-2078, Nov. 2015.

[12] Jason S.Babcock and Jeff B. Pelz, “Building a lightweight eye tracking headgear,” Eye Tracking Research & Application, Texas, 2004, pp.109-113.

[13] Michał Kowalik, “How to build low cost eye tracking glasses for head mounted system,” Faculty of Computer Science and Information Technology, Szczecin Poland, Sep. 2010, pp.1-7.

[14] Zhor Ramdane-Cherif and Amine Naït-Ali, “An Adaptive Algorithm for Eye Gaze Tracking Device Calibration,” IEEE Transactions on Instrumentation and Measurement, vol. 57, no. 4, pp. 716-723, Apr. 2008.

[15] Pengyi Zhang, Zhiliang Wang, Siyi Zheng, and Xuejing Guo, “A Design and Research of Eye Gaze Tracking System Based on Stereovision,”

Proceedings of the 5th international conference on Emerging intelligent computing technology and applications, Berlin, Sep. 2009, pp.278-286.

[16] Krzysztof Murawski, “Method for Determining the Position of the Pupil for Eye Tracking Applications,” Methods and Models in Automation and Robotics, Miedzyzdroje, Aug. 2010, pp.356-362.

[17] Rafael Santos, Nuno Santos, Pedro M. Jorge, and Arnaldo Abrantes, “Eye Gaze as a Human computer Interface,” Conference on Electronics Telecommunications and Computers, Portugal, Nov. 2014, pp.376-383.

[18] Pavol Fabo and Roman Durikovic, “Automated Usability Measurement of Arbitrary Desktop Application with Eyetracking,” International Conference on Information Visualisation, Montpellier, Jul. 2012, pp.625-629.

[19] Huabiao Qin, Jun Liu, and Tianyi Hong, “An eye state identification method based on the Embedded Hidden Markov Model,” Vehicular Electronics and Safety (ICVES), Istanbul, Jul. 2012, pp.255-260.

[20] Bumhwi Kim, Yonghwa Choi, and Minho Lee, “Embedded face recognition system considers human eye gaze using glass-type platform,”

International Conference on Consumer Electronics, Las Vegas, Jan. 2014, pp.357-358.

[21] Luo Chao and Su Jian-Bo, “Eye localization for embedded system,”

Control Conference 31st Chinese, Hefei, Jul. 2012, pp.3828-3833.

[22] Matej Cerny and Martin Dobrovolny, “Eye tracking system on embedded platform,” International Conference Applied Electronics, Pilsen, Sep. 2012, pp.51-54.

[23] Ruian Liu, Shengtao Ma, Mimi Zhang, and Lei Wang, “Implementation and Optimization of the Eye Gaze Tracking System Based on DM642,”

Intelligent Networks and Intelligent Systems, Shenyang, Nov. 2010, pp.48-51.

[24] Tomoaki Ando, Vasily G. Moshnyaga, Koji Hashimoto, “A low-power FPGA implementation of eye tracking,” International Conference on Acoustics, Speech and Signal Processing, Kyoto, Mar. 2012, pp.1573-1576.

[25] Rafael Santos, Nuno Santos, Pedro M. Jorge, and Arnaldo Abrantes, “Eye Gaze as a Human computer Interface,” Conference on Electronics Telecommunications and Computers, Portugal, Nov. 2014, pp.376-383.

[26] Louise H. Crockett, Ross A. Elliot, Martin A. Enderwitz, and Robert W.

Stewart, Embedded Processing with the ARM^® Cortex^®-A9 on the Xilinx^® Zynq^®-7000 All Programmable SoC, Scotland, 2014.

[27] Javier Cerezuela-Mora, Elisa Calvo-Gallego, and Santiago Sanchez-Solano, “Hardware/Software co-design of video processing applications on a reconfigurable platform,” International Conference on Industrial Technology, Seville, Mar. 2015, pp.1694-1699.

[28] M. Ali Altuncu, Taner Guven, Yasar Becerikli, and Suhap Sahin,

“Real-Time System Implementation for Image Processing with Hardware/Software Co-design on the Xilinx Zynq Platform,” International Journal of Information and Electronics Engineering, vol. 5, no. 6, pp.

473-477, Nov. 2015.

[29] Marcos Nieto, Oihana Otaegui, Gorka Velez, Juan Diego Ortega, and Andoni Cortes, “On creating vision based advanced driver assistance systems,” The Institution of Engineering and Technology, vol. 9, no. 1, pp.59-66, Jan. 2015.

[30] Matthew Russell and Scott Fischaber, “OpenCV Based Road Sign Recognition on Zynq,” International Conference on Industrial Informatics, Bochum, Jul. 2013, pp.596-601.

[31] Mauro Turturici, Sergio Saponara, Luca Fanucci, and Emilio Franchi,

“Low-power embedded system for real-time correction of fish-eye automotive cameras,” Design, Automation & Test in Europe Conference &

Exhibition, Dresden, Mar. 2012, pp.340-341.

[32] Michael Hubner, “Design of an Attention Detection System on the Zynq-7000 SoC,” International Conference on ReConFigurable Computing and FPGAs, Cancun, Dec. 2014, pp.1-6.

[33] Roland Dobai and Lukas Sekanina, “Image Filter Evolution on the Xilinx Zynq Platform,” Conference on Adaptive Hardware and Systems, Torino, Jun. 2013, pp.164-171.

[34] Muhammed Al Kadi, Patrick Rudolph, Diana Gohringer, and Michael Hubner, “Dynamic and partial reconfiguration of Zynq 7000 under Linux,”

International Conference on Reconfigurable Computing and FPGAs, Cancun, Dec. 2013, pp.1-5.

[35] TaiLin Han, Hua Cai, GuangwWen Liu, and Bo Wang, “The Face Detection and Location System Based on Zynq,” International Conference on Fuzzy Systems and Knowledge Discovery, Xiamen, Aug. 2014, pp.835-839.

[36] Yang Yuan and Yuehui Tan, and Huixian Sun, “Design of the software framework of reconfigurable communication interface test system based on Zedboard,” Software Engineering and Service Science, Beijing, Sep.

2015, pp.754-757.

[37] 黃奇武、吳友名、葉勇志，“以 ZedBoard 實現螢幕眼動儀”，2015 年 民生電子研討會，彰化、臺灣，2015 年 11 月，139 頁。

[38] Yiu-ming Cheung and Qinmu Peng, “Eye Gaze Tracking With a Web Camera in a Desktop Environment,” IEEE Transactions on Human Machine Systems, vol. 45, no. 4, pp. 419-430, Aug. 2015.