Efficiently tracking and imaging interested moving targets is crucial across various applications, from autonomous systems to surveillance. However, persistent challenges remain in various field

Efficiently tracking and imaging interested moving targets is crucial across various applications, from autonomous systems to surveillance. However, persistent challenges remain in various fields, including environmental intricacies, limitations in perceptual technologies, and privacy considerations. In this study, we present a teacher-student learning model, generative adversarial network (GAN)-guided diffractive neural network (DNN), which performs visual tracking and imaging of the interested moving target. GAN, as a teacher model, empowers efficient acquisition of the skill to differentiate the specific target of interest in the domains of visual tracking and imaging. DNN-based student model learns to master the skill to differentiate the interested target from GAN. The process of obtaining a GAN-guided DNN starts with capturing moving objects effectively by using an event camera with high temporal resolution and low latency. Then the generative power of GAN is utilized to generate data with position-tracking capability for the interested moving target, subsequently serving as labels to the training of DNN. DNN learns to image the target during training while retaining the target's positional information. Our experimental demonstration highlights the efficacy of GAN-guided DNN in visual tracking and imaging the interested moving target. We envision the GAN-guided DNN can significantly enhance autonomous systems, and surveillance.show less

Coherent optics are emerging as promising solutions for future passive optical networks. However, upstream burst-mode coherent detection faces challenges due to the need for fast digital signal

Coherent optics are emerging as promising solutions for future passive optical networks. However, upstream burst-mode coherent detection faces challenges due to the need for fast digital signal processing and its susceptibility to laser wavelength drift. To address these issues, we propose an algorithm capable of rapid channel equalization and frequency offset estimation (FOE). The feasibility of the proposed scheme is experimentally verified through 128-Gbits/s 16QAM signal transmission systems. Consequently, integrating a fine FOE tap into the AEQ allows for rapid convergence and accurate frequency offset estimates within approximately ±0.5 times the symbol rate, while maintaining low complexity.show less

The microring resonator (MRR) plays important roles in signal processing because high quality band-pass filtering can be obtained at its drop port. To promote signal to noise ratio, a high rejec

The microring resonator (MRR) plays important roles in signal processing because high quality band-pass filtering can be obtained at its drop port. To promote signal to noise ratio, a high rejection ratio is significantly demanded. However, it is still challenging to promote the rejection ratio of the MRR-based band-pass filter. To solve this problem, we propose to use an all-pass filter (APF) to enhance the rejection ratio of the MRR-based band-pass filter. Experimental results show that the improved rejection ratio is as high as 47.7 dB, which is improved by 23.6 dB compared with that of the MRR. Meanwhile, the bandwidth of the MRR-based band-pass filter is reduced from 2.61 to 1.14 GHz thanks to the constructive interference in the passband. Additionally, the center frequency of this ultra-high rejection MRR can be continuously tuned from 6.26 to 46.25 GHz. The quality(Q) of the MRR is improved from 7.4×104 to 1.7×105.During the adjustment, the rejection ratio of the band-pass filter remains exceeding 40 dB. The proposed approach can be used to achieve optical bandpass filters with high performance.show less

Spectroscopy, especially for plasma spectroscopy, provided a powerful platform for biological and material analysis with their elemental and molecular fingerprinting capability. AI has the treme

Spectroscopy, especially for plasma spectroscopy, provided a powerful platform for biological and material analysis with their elemental and molecular fingerprinting capability. AI has the tremendous potential to build a universal quantitative framework covering all branches of plasma spectroscopy based on its unmatched representation and generalization ability. Herein, we introduce an AI-based unified method called self-supervised image-spectrum twin information fusion detection (SISTIFD) to collect twin co-occurrence signals of the plasma and to intelligently predict the physical parameters for improving the performances of all plasma spectroscopic techniques. It can fuse the spectra and plasma images in synchronization, derive the plasma parameters and provide accurate results. The experiments demonstrates its excellent utility and capacity with 98% reduction in evaluation indexes and 143 Hz of analysis frequency. Besides, as a completely end-to-end and self-supervised framework, the SISTIFD enables automatic detection without manual preprocessing or intervention. With these advantages, it remarkably enhanced various plasma spectroscopic techniques with state-of-the-art performance and unsealed their possibility in industry, especially in the regions that require both capability and efficiency. This scheme brings new inspiration to the whole field of plasma spectroscopy, and enables in-situ analysis with a real-world scenario of high throughput, cross-interference, various analyte complexity, and diverse applications.show less