Comparing Pose Estimation Models For Lower-Body Joints
Bibliographic record
Abstract
Whole-body pose estimation models are commonly evaluated and compared for all joints. While this provides insight into the performance of pose estimation models, it can be misleading for applications that rely on specific sets of joints, such as gait and mobility assessment, which generally depend on lower-body joints. In this work, we aim to prove that considering all joints for comparison and evaluation of pose estimation models for joint-specific applications can be misleading. Our work highlights that even models with high overall accuracy can show poor performance with joint-specific evaluation, specifically for lower-body joints, as these are more prone to occlusion, poor contrast with the background, and fast motions. This emphasizes the need for evaluation of pose estimation models for lower-body joints for clinical and biomechanical applications and for a targeted methodology to achieve this evaluation. In this work, we propose an approach to evaluate the performance of pose estimation models for lower-body joints necessary for gait and mobility applications. We tested six state-of-the-art human pose estimation models on HumanEva and MoVi datasets. Our results show that models like MediaPipe show superior performance for whole-body evaluation, but underperform when evaluated only for lower-body pose estimation
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.015 |
| Meta-epidemiology (narrow) | 0.004 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.006 | 0.004 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".