FRECUENCIA RESPIRATORIA EN IMÁGENES

Respiratory rate (FR) is the number of breaths per minute performed by a living being and is regulated by the nervous system and some other physiological or pathological characteristics. FR can be extracted using  image photoplethysmography  (IPPG), which is a technique of extracting different physiological variables from a person's video, remotely and contactlessly. For this research it was decided to study FR due to its importance in the clinical context, as it reflects some essential functions of the body and thus evaluate its functioning. It may be useful, for example, to detect respiratory and/or cardiac diseases. In this project, a total of 20 participants were studied, the ages of which ranged from 18 to 50 years. For the acquisition of the videos, the camerasthat each participant had were used, suchas the integrated camera of a computer or the camera of a cell phone. The exercise consisted of recording a video by each participant for about 3 minutes, where he was told how to breathe to control his FR and then compare it with the results. The measurement of FR was done through the application "Paced Breathing" already used in previous research. For the extraction of the signal was first performed the location of a region of interest (ROI) on the face,with the algorithm of Viola-Jones, and then apply an artifact removal stage. Through image processing techniques and statistical processing of multivariate signals, the FR of each subject with an average error of 15% could be extracted. Preprocessing tools such as pass-high and pass-down filters,  linear trend removal, and moving average filters wereused. For the extraction of the IPPG signal and fr the Wavelet transform was used and then a post-processing for the removal of  unwanted components. The method proposed in this work is novel and presentsanalternati goes for the extraction of different physiological variables. Participants' recording conditions are not under many restrictions, so it could be used in various health fields without the need for demanding  conditions. In addition, it shows similar performance to the state of the art and provides a noninvasive way for measuring a physiological variable of clinical and research importance.