Federated Fine-Tuning of Video Segmentation Foundation Models With Shared Long-Term Memory Mechanisms. JDIAS 2026, 1 (1).