Vital signs such as heart rate and heart rate variability can be acquired using a variety of equipment, such as ECG, pulse oximeter and even video camera. ECG using wet electrodes is considered the gold standard, but it is not suitable for longterm patient monitoring. Dry electrodes could solve this problem, but the motion of the sensor relative to the skin affects measurements. Noncontact modalities (e.g. heart rate detected from a video of patient’s face) could offer further advancements in patient care, but again motion artifacts, caused by changing illumination conditions, affect measurements. Mainstream processing techniques typically assume ideal conditions and fail under realistic conditions. This thesis pinpoints the failure mechanisms of a few commonly used heart rate estimation methods under realistic conditions and proposes mitigation techniques, hoping to contribute to the effort of increasing adoption rate of modern and convenient sensor technologies for patient care.