| Abstract: |
Surface Plasmon Resonance have been a gold standard for biosensing and chemical sensing over the past decades. The surface plasmons are a confined electromagnetic wave mode propagating on surface of noble metals. One of the key features of surface plasmons is that they are sensitive to its surrounding medium, therefore the surface plasmons are usually applied in sensing applications. It has been very well established that measuring the phase response of the surface plasmons is more sensitive and more robust compared to intensity or amplitude measurements. To measure the phase, of course, an interferometer is required. This will impose the complexity to the optical alignment. Moreover, the interferometric systems usually require a well-controlled experimental condition, such as, vibration isolation system. Recently, there are some interest of the research community to recover the surface plasmons phase through computational phase retrieval algorithms, such as, Ptychography. Although these computational algorithms can recover the phase profile, they do require many images or a lengthy computing time making them not suitable for real-time measurement. Here, we propose a novel approach to perform surface plasmon phase retrieval using image recognition though deep learning. We demonstrate the feasibility of using the deep learning to recover amplitude and phase responses of simulated back focal plane images. |