Revisiting Reliability in the Reasoning-based Pose Estimation Benchmark
This work addresses benchmark reliability problems for researchers evaluating pose-aware multimodal large language models, but it is incremental as it focuses on annotation refinement rather than new methods.
The paper identified reproducibility and quality issues in the reasoning-based pose estimation benchmark, such as mismatched image indices and inherent limitations like redundancy and imbalance, and released refined ground-truth annotations to improve consistency.
The reasoning-based pose estimation (RPE) benchmark has emerged as a widely adopted evaluation standard for pose-aware multimodal large language models (MLLMs). Despite its significance, we identified critical reproducibility and benchmark-quality issues that hinder fair and consistent quantitative evaluations. Most notably, the benchmark utilizes different image indices from those of the original 3DPW dataset, forcing researchers into tedious and error-prone manual matching processes to obtain accurate ground-truth (GT) annotations for quantitative metrics (\eg, MPJPE, PA-MPJPE). Furthermore, our analysis reveals several inherent benchmark-quality limitations, including significant image redundancy, scenario imbalance, overly simplistic poses, and ambiguous textual descriptions, collectively undermining reliable evaluations across diverse scenarios. To alleviate manual effort and enhance reproducibility, we carefully refined the GT annotations through meticulous visual matching and publicly release these refined annotations as an open-source resource, thereby promoting consistent quantitative evaluations and facilitating future advancements in human pose-aware multimodal reasoning.