本研究從企業選才角度出發，針對企業傳統招募的過程中所使用的面談方式，不易完全了解面試者的全部訊息或人格特徵。
本研究透過人臉辨識及AI機器學習的技術，並結合『面相學』中相關論說，發展出『臉部即時面相特徵分析及建議』平台,提供人潛在的性格分析，在結合面相有職能定義後，期望能提升人員甄選人才的準確度與效率
針對面相學與選才之整合分析，本研究是將擷取的人臉圖像上傳至微軟Azure雲端平台，進行臉部特徵值運算分析，透過特徵值及面相學資料庫比對，最後得到專業的面相建議，在日後持續的樣本分析與AI學習之後，期能提供企業於選才時，能有更準確及效率的判斷依據。

The research starts from the perspective of the selection of enterprises’ talents. While focusing on the process of traditional recruitment, current implementations have always failed to solve the problem of talents recruitment.
In order to understand the limitations of current academics, it’s worthy to explore the advantages and disadvantages of this domain. Through the techniques of facial recognition and machine learning combined with facial features in Physiognomy, we develop a platform called “Instant Personality Analysis of Facial Features”. Through integrating Artificial Intelligence, big data with facial and competency definitions, this platform provides us the information of potential personality analysis, expected to improve recruitment more efficiency, reduce potential risks and make up a deficiency of other selection tools.
This study adopts the methods of general natural science experimentations, and literature review, for analyzing the correlation between Physiognomy and talent selection. The research utilizes open source technology, facial features detection and Cloud System’s instant facial image classification technology to dynamically and immediately capture images, then uploading to the cloud artificial intelligence service of analysis. The facial features will be compared to Physiognomy database thus obtain the facial analysis of the subject. The recommendations obtained from the correlation results of Physiognomy and selection analysis can provide a basis for the company to consider more when selecting talents

參考文獻
Berbar, M. A. (2006). Faces and Facial Features Detection in Color Images. Paper presented at the Geometric Modeling and Imaging--New Trends (GMAI'06).
Holland, J. (1970, 2019/02/04). Personality-Job Fit Theory. Retrieved from https://wiki.mbalib.com/zh-tw/霍兰德职业兴趣理论
McCurrie, M., Beletti, F., Parzianello, L., Westendorp, A., Anthony, S., & Scheirer, W. J. (2018, 2019/02/15). Convolutional Neural Networks for Subjective Face Attributes. Retrieved from https://www.journals.elsevier.com/image-and-vision-computing
Microsoft. (2017). Microsoft Cognitive Services - Face. Retrieved from https://azure.microsoft.com/en-us/services/cognitive-services/face/
Microsoft. (2018, 2019/01/25). Microsoft Cognitive Services - Custom Vision. Retrieved from https://azure.microsoft.com/zh-tw/services/cognitive-services/custom-vision-service/
Rosebrock, A. (2017, 2017/04/3). Facial landmarks with dlib, OpenCV, and Python. Retrieved from https://www.pyimagesearch.com/2017/04/03/facial-landmarks-dlib-opencv-python/
Wang, G. T. (2018, 2019/04/20). Python 使用 OpenCV、Dlib 實作即時人臉偵測程式教學. 一般網站. Retrieved from https://blog.gtwang.org/programming/python-opencv-dlib-face-detection-implementation-tutorial/
WIKI. (2000, 2019/02/02). OpenCV. Retrieved from https://zh.wikipedia.org/wiki/OpenCV#%E5%8F%82%E8%80%83%E6%96%87%E7%8C%AE
Zhou, F. (2007, 2007/08/24). 神奇的職場現象：六種個性分別最適合的職業. Retrieved from https://tw.aboluowang.com/2007/0824/53111.html
壬水漫. (2017, 2017/7/24). 為什麼古代選拔人才首先要看他的面相. 命理風水. Retrieved from https://kknews.cc/zh-tw/history/qjb8myo.html
宮澤美智. (2015). 圖解五官清楚易懂 如何看面相.
彭雍森. (2016). 基於人臉偵測之面相決策系統設計. 健行科技大學, Retrieved from https://hdl.handle.net/11296/9y26kt
心理研究小組. (2010). 人際關係不能不學的面相術.
楊凱麟. (2013). 基於浮動參數與生物特徵的人臉特徵偵測. 朝陽科技大學, Retrieved from https://hdl.handle.net/11296/98qfkc
湯德為. (2008). 面相學運用於台灣地區企業選才之研究. 開南大學, Retrieved from https://hdl.handle.net/11296/8ptkee
潘鈺濬. (2008). 另類的甄才工具－命理. 國立中山大學, Retrieved from https://hdl.handle.net/11296/g85c6v
王麗芬. (1982). 面相術.
羅正漢. (2018, 2018/03/16). 人臉辨識應用全面爆發！AI加持與終端應用是關鍵. Retrieved from https://www.ithome.com.tw/tech/121690
胡銜之. (2015). 基於OpenCV實現搜尋影像人臉辨識系統之研究. 國立彰化師範大學, Retrieved from http://thesis.lib.nccu.edu.tw/record/#G00M0254011%22.
謝沅瑾. (2015). 觀相 謝沅瑾教你知人知面又知心，100張3D面相詳細圖解貴人、小人真面目.
陳勇緯. (2010). 可應用於面相學之即時人臉特徵擷取. 2010 資訊科技國際研討會論文集.
陳哲毅. (2005). 第一次學面相學就做對.
黃恆堉. (2013). 面相一本通.