This research introduces a novel open-vocabulary semantic segmentation approach that improves how assistive robots understand and navigate indoor environments.

• Integrates uncertainty alignment techniques to enhance recognition accuracy in complex indoor settings
• Enables smart wheelchairs and other assistive robots to better identify spatial regions and functional areas
• Bridges the gap between computer vision capabilities and practical mobility solutions for people with disabilities
• Demonstrates real-world applications for autonomous navigation in built environments

This advancement represents a significant step forward for medical assistive technologies by creating more reliable, intuitive mobility solutions for individuals with physical disabilities, potentially increasing independence and quality of life.

Open-Vocabulary Semantic Segmentation with Uncertainty Alignment for Robotic Scene Understanding in Indoor Building Environments