Feeding human keypoints into GPT-4o prompts to create pose-focused instruction data and fine-tuning LLaVA-1.5 with it raises scores on the authors' self-generated E-HPAUB benchmark.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
LLaVA-Pose: Enhancing Human Pose and Action Understanding via Keypoint-Integrated Instruction Tuning
Feeding human keypoints into GPT-4o prompts to create pose-focused instruction data and fine-tuning LLaVA-1.5 with it raises scores on the authors' self-generated E-HPAUB benchmark.