Implementation on Identity RecognitionSystem Using Mobile device and CloudService

• Main Authors: WU,WEI-SHENG, 吳偉生
• Other Authors: HUANG,DENG-YUAN
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/04817885062468516581

碩士 === 大葉大學 === 電機工程學系 === 103 === Computers and cameras are required for the research of image processing and computer vision. In recent years, due to the rapid development of smart handheld devices, they are frequently equipped with cameras, which has a recording function as well as with the hardware specifications comparable with a desktop computer. Based on the statements mentioned earlier, it is not difficult to find that smart handheld devices have the potential to replace desktop computers in computing power. In practice, smart handheld devices may encounter two very big problems when implementing a face recognition system, namely: (1) the establishment of a face database: To implement a face recognition system, it is required to build a very large face database for subsequent training and matching. However, due to limited storage space of smart handheld devices, it can cause fail to successfully implement a face recognition system on them, and (2) the heavy computational cost: Because face recognition algorithm is computationally intensive, it is easily beyond the capability of smart handheld devices’ hardware limitation, resulting in shutdown of themselves or fail operations of Apps. For these reasons, this thesis presents a solution to combine smart handheld devices and cloud services for realizing a face recognition system. Experiments show that the proposed method can realize the tasks of image transmission, facial feature extractions, and face recognition in dozens of milliseconds. Moreover, the face recognition rate can reach a satisfactory level, indicating the feasibility of the proposed method.