IDMR introduces a new paradigm for precise instance-level visual correspondence that significantly enhances object tracking across varied scenarios.

• Enables retrieval of identical objects/instances within different contexts based on both visual and textual queries
• Overcomes limitations of current multimodal systems that struggle with instance-level precision
• Demonstrates practical applications in security monitoring and surveillance systems
• Provides a foundation for advanced object tracking in complex, real-world environments

This research is particularly valuable for security applications where tracking specific objects or individuals across multiple scenarios and camera feeds is critical for effective monitoring and threat detection.

IDMR: Towards Instance-Driven Precise Visual Correspondence in Multimodal Retrieval