Microwave transversal filters, which are implemented based on the transversal filter structure in digital signal processing, offer a high reconfigurability for achieving a variety of signal processing functions without changing hardware. When implemented using microwave photonic (MWP) technologies, also known as MWP transversal filters, they provide competitive advantages over their electrical counterparts, such as low loss, large operation bandwidth, and strong immunity to electromagnetic interference. Recent advances in high-performance optical microcombs provide compact and powerful multi-wavelength sources for MWP transversal filters that require a larger number of wavelength channels to achieve high performance, allowing for the demonstration of a diverse range of filter functions with improved performance and new features. Here, we present a comprehensive performance analysis for microcomb-based MWP spectral filters based on the transversal filter approach. First, we investigate the theoretical limitations in the filter spectral response induced by finite tap numbers. Next, we analyze the distortions in the filter spectral response resulting from experimental error sources. Finally, we assess the influence of input signal’s bandwidth on the filtering errors. These results provide a valuable guide for the design and optimization of microcomb-based MWP transversal filters for a variety of applications.
Deep learning is a powerful tool for image analysis and medical applications. However, due to their intricate black-box nature, comprehending deep learning model predictions are often challenging. Oral cancer is globally prevalent, necessitating reliable AI algorithms for screening, especially for low-income regions. Interpretability is crucial for reliable AI. Visual explanation, generating attention maps highlighting decision-influencing regions, aids interpretability and also helps guide AI focus. Elevating AI reliability involves assessing decision confidence as well. Quantifying model output certainty helps identify uncertain cases, which need additional examination. Dataset quality is also pivotal for reliable AI development. Methods to evaluate and enhance the data and label/annotation quality will also be essential.
Oral cancer is a global public health challenge, particularly affecting low-resource regions. To address the early detection requirement, we introduce a novel intraoral probe combining conventional oral examination (COE), autofluorescence visualization (AFV), and optical coherence tomography (OCT) for multidimensional oral cancer screening. Real-time COE and AFV offer a broad field of view, while OCT provides depth-resolved imaging. Our handheld probe demonstrates widefield, autofluorescence, and depth-resolved imaging capabilities in clinical settings, holding promise for enhanced early detection and management of oral cancer.
