Gait Analysis Using Computer Vision Based on Cloud Platform and Mobile Device

Table 1

Results of the sagittal HS and TO detection algorithm showing the amount of correct detections (less than 2 frames of difference between algorithm and manual marking), undetected cases, wrong detection (more than 2 frames of difference), and the root mean square error of both correct and wrong cases.

Approach	Correct	Undetected	Wrong	RMSE
DAI dataset normal gait heel strike
Sagittal	90.2%	1.1%	7.6%	1.44 frames (48 ms)
Frontal	89.4%	0%	10.60%	1.88 frames (63 ms)
Toe off
Sagittal	93.3%	2.2%	2.2%	1.08 frames (36 ms)
Frontal	89.4%	0%	10.6%	1.63 frames (54 ms)
DAI dataset abnormal gait heel strike
Sagittal	89%	2.1%	6.9%	1.79 frames (60 ms)
Frontal	72.1%	0%	27.9%	2.42 frames (81 ms)
Toe off
Sagittal	82.1%	3.6%	10.7%	1.59 frames (53 ms)
Frontal	75%	0%	25%	2.17 frames (72 ms)
Total heel strike
Sagittal	89.5%	1.7%	7.2%	1.66 frames (55 ms)
Frontal	78.8%	0%	21.3%	2.23 frames (74 ms)
Toe off
Sagittal	86.5%	3.0%	7.4%	1.41 frames (47 ms)
Frontal	80.6%	0%	19.4%	1.98 frames (66 ms)