我們提出一個應用深度學習架構學習人臉辨識並進行即時心跳量測的方法。本方法實驗結果可以得知，OpenCV自帶人臉辨識與YOLOv2實作出來之即時心率測量之差異，可以比較出OpenCV自帶人臉辨識速度及辨識率不佳，使用YOLOv2來作人臉辨識後，能明顯提升運行效果與準確度，因為YOLOv2不同OpenCV自帶人臉辨識的方法，OpenCV自帶人臉重複搜尋整張圖的相關特徵點，而YOLOv2只對整張圖有興趣的特徵點進行人臉搜索，由此我們可以因為加快的運行速度，而達到未來需要追加其他功時，能擁有更高的拓展可能性。
We propose a method of applying deep learning architecture to learn face recognition and perform instant heartbeat measurement. The experimental results show that OpenCV comes with people. The difference between face recognition and real heart rate measurement made by YOLOv2 can compare the speed and recognition rate of OpenCV's own face recognition. After using YOLOv2 for face recognition, it can significantly improve the running effect and accuracy, because YOLOv2 Different OpenCV's own face recognition method, OpenCV comes with a face to repeatedly search for the relevant feature points of the whole picture, and YOLOv2 only performs face search on the feature points that are of interest to the whole picture, so we can run faster. Speed, and when you need to add other work in the future, you can have a higher expansion potential.
