Automatic Image Annotation (AIA) plays an important role in large-scaled intelligent image management and retrieval. Based on the correlation between image low-level features and high-level semantic concepts, images can be efficiently retrieved from large-scaled image dataset. Recently, many researchers leverage machine learning techniques to annotate images automatically. However, these methods still have many challenges regarding efficiency and scalability in the massive image dataset. Moreover, manually labeling massive images is a costly and time-consuming task, which is unacceptable in practical applications. Therefore, only a few labeled images can be obtained as samples in the training dataset. However, the tags associated with labeled and unlabeled images found on social network websites may be helpful for improving the performance of AIA. In this work, we propose a Multi-Modal Semantic Hash Learning framework named MMSHL for AIA. MMSHL seamlessly integrates multi-graph learning, multi-modal correlation learning and latent semantic hashing learning into a joint optimization framework. Based on MMSHL, we annotate images using a two-step semi-supervised learning approach. Since our AIA method makes use of associated tags of images, good results can be achieved. Extensive experiments are performed based on two real-world datasets MIR Flickr and NUS-WIDE. Experimental results show that our framework can improve the performance of AIA effectively.

In scenes with low lighting at night, the outline of objects that need to be recognized, such as vehicles, is not clear, and cannot be accurately recognized by the automatic driving system. At present, there are many researches on instance segmentation models, but there are few researches on the instance segmentation application of automatic driving night scenes. According to BDD100K dataset, the automatic driving daytime scene dataset is marked. First, we perform data augmentation by using gamma correction to simulate the night driving scene in the training phase. Then we use our improved low-light enhancement algorithm with gradient increment based on RetinexNet in the prediction phase to brighten night driving scene images. Furthermore, we evaluated our proposed method on YOLACT++ model. The results show that the improved YOLACT++ automatic driving night segmentation ability has been significantly improved, the segmentation of vehicles at night is more accurate and robust, and it has better application value in night automatic driving scenarios.

The dispersion measures (DMs) of the fast radio bursts (FRBs) are a valuable tool to probe the baryonic content of the intergalactic medium and the circumgalactic medium of the intervening galaxies along the sightlines. However, interpreting the DMs is complicated by the contribution of the hot gas in and around our Milky Way. This study examines the relationship between $D{M}_{\mathrm{MW}}$, derived from localized FRBs, and the Galaxy’s hot gas, using X-ray absorption and emission data from O vii and O viii. We find evidence for a positive correlation between $D{M}_{\mathrm{MW}}$ and O vii absorption, reflecting contributions from both the disk and halo components. This conclusion is supported by two lines of evidence: (1) no correlation between $D{M}_{\mathrm{MW}}$ and O vii/O viii emission, which primarily traces dense disk regions; and (2) the comparison with electron density models, where $D{M}_{\mathrm{MW}}$ aligns with models that incorporate both disk and halo components but significantly exceeds predictions from pure disk-only models, emphasizing the halo’s role. Furthermore, the lack of correlation with O viii absorption suggests that the primary temperature of the Galaxy’s hot gas is likely around $2\times {10}^6$ K or less, as traced by O vii absorption, while gas at higher temperatures (∼3–5 × ${10}^6$ K) is present but less abundant. Our findings provide insights into the Milky Way’s gas distribution and improve $D{M}_{\mathrm{MW}}$ estimates for future cosmological studies.

Fast radio bursts (FRBs) provide sensitive probes of ionized baryons through their dispersion measures (DMs). In this work, we test whether localized FRB sight lines intersecting foreground galaxy groups exhibit a measurable excess in cosmic DM relative to clean nonintersecting sight lines. We crossmatch 117 localized FRBs with measured DMs and host redshifts against a foreground group catalog from the DESI Legacy Imaging Surveys. Within the catalog footprint and at z _host < 0.5, we identify 23 sight lines intersecting foreground groups and 18 clean nonintersecting sight lines for the main comparison. A joint fit to this sample gives a positive offset of 143 . 3 − 53.0 + 53.2 pc cm − 3 in the inferred cosmic DM, DM _cosmic , for the group-intersecting sight lines relative to the nonintersecting control sample. A template fit for the foreground halos gives A halo = 0.3 4 − 0.50 + 0.51 , with a positive posterior median but a weakly constrained detailed halo contribution. A mock forecast within the DESI footprint indicates that larger samples of localized FRBs would improve the foreground halo-template constraint.