Real Time Powerlifting Form Assessment using Yolov5 and Mediapipe
Bibliographic record
Abstract
This study introduces an on-device, real-time AI posture correction system tailored for the three corepowerlifting exercises: bench press, back squat, and conventional deadlift. We employ YOLOv5 for efficient persondetection, combined with an HSV-based background subtraction mask to dynamically crop the region of interest(ROI), ensuring streamlined processing. To enable accurate, user-tailored feedback, we present the PolyViewKinematic Corpus (PKC), a novel multi-camera 3D landmark dataset recorded from front, side, and oblique angles,capturing concentric and eccentric phases from 150 lifters with diverse body types. Utilizing PKC data, we train onlymachine learning classification algorithms (e.g., decision trees, LightGBM) to detect subtle joint alignment deviationsat rep “bottom” and “lockout” phases, with evaluations showing these methods achieve high accuracy in lift phaseclassification. Our phase-sensitive feedback system, based on user-calibrated angle bands and temporal smoothing,delivers precise visual overlays and concise audio prompts to guide lifters toward safer, more effective techniques.Deployed via a low-latency web interface, the system provides exercise-specific cues in under 50 ms per frame, helpinglifters minimize injury risk and optimize performance without external sensors or complex setup.
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 | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.007 | 0.002 |
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".