The ability to detect and image in the near-infrared (NIR) regions, specifically NIR-I (700-900 nm) and NIR-II (1000-1700 nm), is crucial for medical diagnostics and bioimaging due to their exceptional tissue penetration and absorption characteristics. Organic photodetectors (OPDs) present a promising solution for self-aligned imaging in these regions, due to their tunable optoelectronic properties and flexible fabrication processes. However, achieving high-performance dual-mode narrowband OPDs for NIR-I and NIR-II detection remains a significant challenge due to the inherently broadband nature of most organic semiconductor materials. In this study, we introduce a bias-switchable dual-mode narrowband OPD with a back-to-back stacked charge collection narrowing/photodiode-type double bulk heterojunction (BHJ) architecture. The dual-mode narrowband OPD exhibits excellent wavelength selectivity and enhanced imaging capabilities, facilitating self-aligned imaging in both NIR-I and NIR-II regions. Additionally, the double BHJ architecture provides design flexibility, allowing for the realization of dual-mode narrowband responses through various BHJ configurations.

A rapid, reliable and nondestructive method for the gender discrimination of silkworm cocoons is of great importance for the production of high-quality silk by the mulberry silkworm industry. This study aimed to determine the feasibility of using soft X-ray imaging with multivariate data analysis to discriminate the gender of silkworm cocoons. X-ray images of silkworm cocoons were obtained and preprocessed and then the region of interest of the chrysalises was segmented. In total, 11 morphological characters of the chrysalises were extracted and compressed by principal component analysis to visualize cluster trends. In developing the discrimination classifiers, four different types of algorithm, including the K nearest neighbors, linear discriminant analysis, back propagation artificial neural network and support vector machine algorithms, were used comparatively; the performances of these models were optimized by cross-validation. The results indicate that the correct identification rates of these classifiers were all high and ranged from 92.57% (obtained via the support vector machine algorithm) to 93.68% (obtained via the K nearest neighbor algorithm). With respect to the running time, the linear discriminant analysis classifier had the best accuracy of 93.31%. These results indicate that this X-ray imaging technique with multivariate analysis could be used successfully for the screening of the gender of silkworm cocoons in the mulberry silkworm industry.