http://repub.eur.nl/res/aut/6256/ List of Publications en http://repub.eur.nl/static-eur/img/logo.png http://repub.eur.nl http://repub.eur.nl/res/pub/10031/ 2007-05-09T00:00:00Z The aim of this thesis was to investigate the feasability of contactless imaging pulse oximetry (proposed term: pulse oxigraphy). The patent disclosed in chapter 2 claims that such pulse oxigraphy can be achieved with camera-derived photoplethysmographic pulse waves at three wavelengths, preferably being 660, 810 and 940nm. From the absorption curves of hemoglobin and oxyhemoglobin it can be easily derived that two of these wavelengths (660 and 940nm) contain oxygenation-related information, and they have proven to be useful for conventional pulse oximetry (in transmission- mode as well as in reflectance-mode). The additional third wavelength (810nm) lies at a so-called isobestic point where the absorption curves of hemoglobin and oxyhemoglobin intersect. Thus, images and/or plethysmographic pulse waves recorded at 810nm do not contain oxygenation-related information, which is useful for reference purposes when dealing with shadows, reflections, movement artifacts and variations in geometry. With regard to pulse oxigraphy the following results were obtained: In chapter 3 we proved that it is possible to derive photoplethysmographic pulse waves containing the heart rythm of a living person at all three required wavelengths from camera recordings collected at a distance of 72 cm. To investigate and validate the capabilities for pulse oxigraphy with this set up, direct measurements on volunteers were sub optimal, because of: Signal-to-noise issues, sequentially recorded heartbeats for oxygen saturation calculations, and lack of a method to induce prolonged stable and adjustable oxygen saturation levels.