AI人臉辨識

Abstract

中文摘要 隨著人工智慧的發展越趨進步，它也逐漸融入我們的生活當中。其中人臉辨識就是其中一項應用，可以應用在門禁管理，身分核對上等。雖然人工智慧在模型耗時的建立較久，但是有了模型便能快速地對影像進行辨識，藉以達到即時的功能。在本專題我們針對人臉辨識這項主題去進行實作與探討。我們利用鏡頭進行影像擷取，再將擷取到的影像藉由人工智慧建立的模型進行辨識。在辨識神經網路模型我們分別使用特徵學習以及深度學習兩種方法建立，並比較兩者結果的差異。最終我們能夠實現實時的人臉辨識系統。

Abstract With the progress of artificial intelligence (AI), it gradually integrated into our life. Face recognition is one of AI applications. It can be applied in access control management, identification, and so on. Although the construction of an AI model is time-consuming, the image can be recognized quickly with the model to achieve real-time function. In this project, we have implemented face recognition by AI. We used the lens to capture images, and then recognized images with AI. In the artificial neural network, we used feature learning and deep learning, respectively, for face recognition. The difference of the two methods was also discussed. Finally, we have demonstrated a real-time face recognition system.