Graph Neural Network Fusion vs. Encoder-Decoder Frameworks in High-Parallax VSLAM Robustness

Robot vision has greatly benefited from advancements in multimodal fusion techniques and vision-language models (VLMs). We adopt a task-oriented perspective to systematically review the applications and advancements of multimodal fusion methods and VLMs in the field of robot vision. For semantic scene understanding tasks, we categorize fusion approaches into encoder-decoder frameworks, attention-based architectures, and graph neural networks. Meanwhile, we also analyze the architectural characteristics and practical implementations of these fusion strategies in key tasks such as simultaneous l

Research goal: How do graph neural network-based fusion approaches compare to encoder-decoder frameworks in terms of robustness to high-parallax conditions in multimodal VSLAM systems, evaluated using the HPatches dataset?

Autonomous synthesis report generated by SOVEREIGN Research Kernel. Tribunal consensus score: 7.7/10.