TY  - RPRT
TI  - Task-Oriented Visual Feature Compression via Residual Vector Quantization for Device-Edge Multimodal Inference
AU  - Luning Pang
AU  - Cheng Yuan
AU  - Jiawei Shao
AU  - Mingtao Huang
AU  - Yuan Shen
PY  - 2026
UR  - https://arxiv.org/abs/2609.37090
ID  - 2609.37090
ER  -