Deep learning approach to estimate foot pressure distribution in walking with application for a cost-effective insole system.
Frederick Mun, Ahnryul Choi
PMID 35034658WHAT IT FOUND
A deep learning model estimated full-foot pressure from just nine sensors in young and older adults walking comfortably.
It outperformed a fuzzy logic model and worked with a low-cost insole prototype, though accuracy was lower for older adults.
Key findings
01The LSTM deep learning model estimated foot pressure distribution from nine sensor inputs with an average relative error of 7.9%, significantly better than the 14.5% error of a traditional fuzzy logic model.
02When tested with a custom low-cost insole prototype, the model predicted pressure with correlation coefficients between 0.63 and 0.97, showing better accuracy for young adults than for older adults.
03The model was trained and validated only on healthy adults walking at a comfortable speed, not on patients with foot pathology or gait disorders.
STILL TO COME
How it was doneWhat they foundWhat it means for PTsWhat it means for OTs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The model was trained and tested only on healthy males with no musculoskeletal disorders. It has not been validated on patients with foot pathology, diabetes, or neurological conditions. Data collection was limited to straight-line walking at a comfortable speed. It is unknown how the model performs during turning, uneven terrain, or different gait speeds. Accuracy was lower for older adults compared to young adults, and the study did not include females, limiting generalizability. The low-cost prototype validation showed wider error margins (up to 12.7% relative error) compared to the commercial system training data.
Declared interests
The study was funded by the Ministry of Education. No other conflicts of interest were declared.
The easy way to misread this
Do not assume this system can currently diagnose foot problems in patients with diabetes or gait disorders. The model was trained exclusively on healthy adults walking in a straight line, so its predictions for pathological gait or clinical populations are unverified.