Graph neural networks are graph-based deep learning technologies that have attracted significant attention from researchers because of their powerful performance. Heterogeneous graph-based graph neural networks focus on the heterogeneity of the nodes and links in a graph. This is more effective at preserving semantic knowledge when representing data interactions in real-world graph structures. Unfortunately, most heterogeneous graph neural networks tend to transform heterogeneous graphs into homogeneous graphs when using meta-paths for representation learning. This paper therefore presents a novel motif-based hierarchical heterogeneous graph attention network algorithm, MBHAN, that addresses this problem by incorporating a hierarchical dual attention mechanism at the node-level and motif-level. Node-level attention aims to learn the importance between a node and its neighboring nodes within its corresponding motif. Motif-level attention is capable of learning the importance of different motifs in the heterogeneous graph. In view of the different vector space features of different types of nodes in heterogeneous graphs, MBHAN also aggregates the features of different types of nodes, so that they can jointly participate in downstream tasks after passing through segregated independent shallow neural networks. MBHAN’s superior network representation learning capability has been validated by extensive experiments on two real-world datasets.

Shui manuscripts are part of the national intangible cultural heritage of China. Owing to the particularity of text reading, the level of informatization and intelligence in the protection of Shui manuscript culture is not adequate. To address this issue, this study created Shuishu_C, the largest image dataset of Shui manuscript characters that has been reported. Furthermore, after extensive experimental validation, we proposed ShuiNet-A, a lightweight artificial neural network model based on the attention mechanism, which combines channel and spatial dimensions to extract key features and finally recognize Shui manuscript characters. The effectiveness and stability of ShuiNet-A were verified through multiple sets of experiments. Our results showed that, on the Shui manuscript dataset with 113categories, the accuracy of ShuiNet-A was 99.8%, which is 1.5% higher than those of similar studies. The proposed model could contribute to the classification accuracy and protection of ancient Shui manuscript characters.

Shui manuscripts provide a historical testimony of the national identity and spirit of the Shui people. In response to the lack of a high-quality Shui manuscripts dataset, we collected Shui manuscript images in the Shui area and used various methods to enhance them. Through our efforts, we created a well-labeled and sizable Shui manuscripts dataset, named Shuishu_T, which is the largest of its kind. Then, we applied target detection technology for Shui manuscript characters recognition. Specifically, we compared the advantages and disadvantages of Faster R-CNN, you only look once (YOLO), and single shot multibox detector (SSD), and subsequently chose Faster R-CNN to detect and recognize Shui manuscript characters. We trained and tested 111 classes of Shui manuscript characters with Faster R-CNN and achieved an average recognition rate of 87.8%. Finally, we designed a WeChat applet that can be used to quickly identify Shui manuscript characters in images obtained by scanning Shui manuscripts with a mobile phone. This work provides a basis for realizing the recognition of characters in Shui manuscripts on mobile terminals. Our research enables the intangible cultural heritage of the Shui people to be preserved, promoted, and shared, which is of great significance for the conservation and inheritance of Shui manuscripts.

Images captured in low-light conditions often have poor visibility, low contrast, and color distortion due to uneven lighting. Most existing enhancement methods often suffer from unstable brightness recovery and color cast, which affect both visual quality and performance of advanced vision tasks. To address those issues, we propose DADNet, a dual-branch network with an attention mechanism and dark channel prior containing an Illumination Enhancement Module (IEM) and Color Transformation Module (CTM). The IEM extracts multi-scale features and improves lighting based on the dark channel prior, while the CTM employs the attention mechanism to handle color features and adjust saturation adaptively. Experimental results on three datasets show that DADNet performs well in both qualitative and quantitative evaluations. It effectively preserves image structure and texture details while achieving a good balance between overall brightness and color quality.

Low-light images often suffer from insufficient illumination and noise, but enhancing dark regions can further amplify noise and degrade texture and structural details. To address these challenges, we propose Zero-MS, a zero-shot method for low-light image enhancement. Zero-MS constructs a noise prior through normalized four-directional cumulative scanning and adaptive selective gating. It then restores illumination through multi-scale information modeling and refines the result using structural and noise responses. Experimental results show that Zero-MS balances information fidelity, visual quality, and noise suppression. The enhanced images exhibit improved brightness, color preservation, and noise suppression, while Zero-MS maintains a lightweight model design.