Table 3.

Oxygen saturation results through leave-one-out. For each subject, the mean of the gradients of the lines of best fit to the data from all other subjects is used as the gradient β of the linear relationship between the data and the oxygen saturation. The intercept c is instead found by using the first minute of data from the subject. The results indicate that the method can not only be used to track oxygen saturation changes, but effectively measure it based on a training data across similar skin colours and under ideal conditions.

	Subject 1	Subject 2	Subject 3	Subject 4	Subject 5
Gradient	39.6	38.5	39.1	38.1	41.9
MAE (%)	2.37	1.76	1.95	2.34	2.71
ME (%)	0.20	0.27	−0.56	0.47	0.45