In this study, the influence of speckle size on contrast-to-noise ratio (CNR) and resolution is examined based on the object dimensions in the macroscopic and microscopic regimes. This research shows that for microscopic samples the conventional scaling laws are no longer effective and the CNR does not counter-propagate in the same manner as the resolution. To our knowledge, a deviation in CNR scaling on speckle size is observed for the first time in the field of microscopic ghost imaging. This result was verified using two different sample shapes. In addition, numerical analysis revealed that the noise of the photodiode is a limiting factor for the CNR. Based on these findings, the conditions for identifying the parameter set that maximizes the CNR and provides high resolution images was defined, which achieving high-quality microscopic ghost images.

Graphene-like two-dimensional (2D) materials have shown remarkable broadband saturable absorption properties. These materials were successfully applied into mode locked lasers to generate laser pulses with the pulse duration from picosecond to femtosecond. However, these novel materials have not shown good performance as far in another important aspect: Q-switched lasers. Solid-state or fiber lasers Q-switched with broadband absorbers usually generated pulses of one hundred nanosecond to several microsecond, which show weak competitiveness compared to traditional absorbers such as Cr: YAG and semiconductor saturable absorption mirror (SESAM). In this paper we utilized BP, WS₂ and MoS₂ solutions as saturable absorbers (SAs) to construct the passively Q-switched Nd:YVO₄ lasers. The pulse durations as short as 2.86 nanosecond was obtained. To the best of our knowledge, it was the first report that the pulse durations approached several nanosecond level in Q-switched lasers with liquid-form of BP, WS₂ and MoS₂ SAs.

Zero-shot learning (ZSL) enables the recognition of novel classes by leveraging semantic knowledge transfer from known to unknown categories. This knowledge, typically encapsulated in attribute descriptions, aids in identifying class-specific visual features, thus facilitating visual-semantic alignment and improving ZSL performance. However, real-world challenges such as distribution imbalances and attribute co-occurrence among instances often hinder the discernment of local variances in images, a problem exacerbated by the scarcity of fine-grained, region-specific attribute annotations. Moreover, the variability in visual presentation within categories can also skew attribute-category associations. In response, we propose a bidirectional cross-modal ZSL approach CREST. It begins by extracting representations for attribute and visual localization and employs Evidential Deep Learning (EDL) to measure underlying epistemic uncertainty, thereby enhancing the model's resilience against hard negatives. CREST incorporates dual learning pathways, focusing on both visual-category and attribute-category alignments, to ensure robust correlation between latent and observable spaces. Moreover, we introduce an uncertainty-informed cross-modal fusion technique to refine visual-attribute inference. Extensive experiments demonstrate our model's effectiveness and unique explainability across multiple datasets. Our code and data are available at: https://github.com/JethroJames/CREST

Summary: High‐resolution imaging is an important issue in various fields of scientific researches and engineering applications. Pseudothermal ghost imaging is one of the subfields of quantum imaging, providing new capabilities beyond conventional imaging methods. Also, it can provide a new viewpoint of imaging physical mechanisms. In this review, we explain the major ideas of pseudothermal ghost imaging, restricting the very important case of high‐resolution imaging. We analyse the strategies which can significantly improve the image quality in pseudothermal ghost imaging. It may apply for merging it with common optical imaging methods in the extreme ultraviolet (XUV) or X‐ray spectral regime for driving the applications to a wider audience in bioscience and nano‐physics.

In this study, we propose a ghost imaging method capable of penetrating dynamic scattering media through a multi-polarization fusion mutual supervision network (MPFNet). The MPFNet effectively processes one-dimensional light intensity signals collected under both linear and circular polarization illumination. By employing a multi-branch fusion architecture, the network excels at extracting multi-scale features and capturing contextual information. Additionally, a multi-branch spatial-channel cross-attention module optimizes the fusion of multi-branch feature information between the encoder and the decoder. This synergistic fusion of reconstruction results from both polarization states yields reconstructed object images with significantly enhanced fidelity compared to ground truth. Moreover, leveraging the underlying physical model and utilizing the collected one-dimensional light intensity signal as the supervisory labels, our method obviates the need for pre-training, ensuring robust performance even in challenging, highly scattering environments. Extensive experiments conducted on free-space and underwater environments have demonstrated that the proposed method holds significant promise for advancing high-quality ghost imaging through dynamic scattering media.