Software Process
The software process (depicted in Figure 1) consists of 7 steps shown in the upper portion of the figure. The process begins with the acquisition of 15s of externally recorded audio data from the patient's trachea (STEP 1). During this time, the patient is instructed to breathe deeply several times throughout the 15s of recording. After this data is recorded, the signal is band-passed at 3 frequency bands (STEP 2), producing three distinct band passed voltage time courses. The relative power in each of these bands is determined (STEP 3) and then zero-meaned. In STEP 7, using linear discriminant analysis (LDA), these values are used to compute the percentage likelihood that the individual has pneumonia, which is output to LED displays on the device.
For more information on the software logic, use the link below.