MambaDFuse: A Mamba-based Dual-phase Model for Multi-modality Image Fusion

Multi-modality image fusion (MMIF) aims to integrate complementary information from different modalities into a single fused image to represent the imaging scene and facilitate downstream visual tasks comprehensively. In recent years, significant progress has been made in MMIF tasks due to advances in deep neural networks. However, existing methods cannot effectively and efficiently extract modality-specific and modality-fused features constrained by the inherent local reductive bias (CNN) or quadratic computational complexity (Transformers). To overcome this issue, we propose a Mamba-based Dual-phase Fusion (MambaDFuse) model. Firstly, a dual-level feature extractor is designed to capture long-range features from single-modality images by extracting low and high-level features from CNN and Mamba blocks. Then, a dual-phase feature fusion module is proposed to obtain fusion features that combine complementary information from different modalities. It uses the channel exchange method for shallow fusion and the enhanced Multi-modal Mamba (M3) blocks for deep fusion. Finally, the fused image reconstruction module utilizes the inverse transformation of the feature extraction to generate the fused result. Through extensive experiments, our approach achieves promising fusion results in infrared-visible image fusion and medical image fusion. Additionally, in a unified benchmark, MambaDFuse has also demonstrated improved performance in downstream tasks such as object detection. Code with checkpoints will be available after the peer-review process.

Research on Joint Representation Learning Methods for Entity Neighborhood Information and Description Information

To address the issue of poor embedding performance in the knowledge graph of a programming design course, a joint represen-tation learning model that combines entity neighborhood infor-mation and description information is proposed. Firstly, a graph at-tention network is employed to obtain the features of entity neigh-boring nodes, incorporating relationship features to enrich the structural information. Next, the BERT-WWM model is utilized in conjunction with attention mechanisms to obtain the representation of entity description information. Finally, the final entity vector representation is obtained by combining the vector representations of entity neighborhood information and description information. Experimental results demonstrate that the proposed model achieves favorable performance on the knowledge graph dataset of the pro-gramming design course, outperforming other baseline models.