TY  - RPRT
TI  - HORNet: Task-Guided Frame Selection for Video Question Answering with Vision-Language Models
AU  - Xiangyu Bai
AU  - Bishoy Galoaa
AU  - Sarah Ostadabbas
PY  - 2026
UR  - https://arxiv.org/abs/2603.18850
ID  - 2603.18850
ER  -