This paper presents a novel, automated, and motion-tolerant method for non-contact cardiac pulse measurement using video imaging and blind source separation. The method uses color video recordings of the human face and applies automatic face tracking along with blind source separation of the color channels into independent components. The approach was validated against an FDA-approved finger blood volume pulse (BVP) sensor and achieved high accuracy and correlation even in the presence of motion artifacts. The technique was also applied to perform heart rate measurements from three participants simultaneously, demonstrating the ability to measure multiple people at the same time. The method uses independent component analysis (ICA) to separate the mixed signals from the RGB color channels. The ICA model assumes that the observed signals are linear mixtures of the underlying sources, and the goal is to find a separating matrix that estimates the original source signals. The technique was tested on 12 participants, with videos recorded using a basic webcam. The results showed that the method achieved high agreement with BVP measurements when participants were sitting still and also in the presence of motion artifacts. The method was able to recover the cardiovascular pulse wave from the video recordings, even when the raw traces did not show clear plethysmographic information. The method was also tested during motion, where participants were allowed to move their heads and bodies slowly while seated. The results showed that the method was robust to motion artifacts and could accurately measure heart rate even when the participant was moving. The method was able to recover the pulse frequency from the video recordings, even when the raw traces were noisy and did not show clear peaks. The method was also tested for simultaneous heart rate measurements of multiple participants, demonstrating the ability to measure multiple people at the same time. The results showed that the method achieved high accuracy and agreement with BVP measurements, even when participants were moving. The method was able to recover the cardiovascular pulse wave from the video recordings, even when the raw traces did not show clear plethysmographic information. The method was also tested for low-light conditions, where the performance of the technique did not vary significantly within the range of luminance used in the experiments. However, the signal-to-noise ratio of the recovered plethysmographic signal was expected to decrease in dim light. The method was also tested for face detection, where the pre-trained frontal face OpenCV classifier was used to detect faces in the video recordings. The results showed that the method was able to detect faces accurately, even in the presence of motion artifacts. The method was also tested for the identification of the ICA component containing the strongest PPG signal. The second component typically contained the strongest pulse signal, but it was not always the case. The method was able to recover the cardiovascular pulse wave from the video recordings, even when the raw traces did not show clear plethysmographic information. The method was also tested for the identification of the ICA component containing the strongest PPG signal. The second component typically containedThis paper presents a novel, automated, and motion-tolerant method for non-contact cardiac pulse measurement using video imaging and blind source separation. The method uses color video recordings of the human face and applies automatic face tracking along with blind source separation of the color channels into independent components. The approach was validated against an FDA-approved finger blood volume pulse (BVP) sensor and achieved high accuracy and correlation even in the presence of motion artifacts. The technique was also applied to perform heart rate measurements from three participants simultaneously, demonstrating the ability to measure multiple people at the same time. The method uses independent component analysis (ICA) to separate the mixed signals from the RGB color channels. The ICA model assumes that the observed signals are linear mixtures of the underlying sources, and the goal is to find a separating matrix that estimates the original source signals. The technique was tested on 12 participants, with videos recorded using a basic webcam. The results showed that the method achieved high agreement with BVP measurements when participants were sitting still and also in the presence of motion artifacts. The method was able to recover the cardiovascular pulse wave from the video recordings, even when the raw traces did not show clear plethysmographic information. The method was also tested during motion, where participants were allowed to move their heads and bodies slowly while seated. The results showed that the method was robust to motion artifacts and could accurately measure heart rate even when the participant was moving. The method was able to recover the pulse frequency from the video recordings, even when the raw traces were noisy and did not show clear peaks. The method was also tested for simultaneous heart rate measurements of multiple participants, demonstrating the ability to measure multiple people at the same time. The results showed that the method achieved high accuracy and agreement with BVP measurements, even when participants were moving. The method was able to recover the cardiovascular pulse wave from the video recordings, even when the raw traces did not show clear plethysmographic information. The method was also tested for low-light conditions, where the performance of the technique did not vary significantly within the range of luminance used in the experiments. However, the signal-to-noise ratio of the recovered plethysmographic signal was expected to decrease in dim light. The method was also tested for face detection, where the pre-trained frontal face OpenCV classifier was used to detect faces in the video recordings. The results showed that the method was able to detect faces accurately, even in the presence of motion artifacts. The method was also tested for the identification of the ICA component containing the strongest PPG signal. The second component typically contained the strongest pulse signal, but it was not always the case. The method was able to recover the cardiovascular pulse wave from the video recordings, even when the raw traces did not show clear plethysmographic information. The method was also tested for the identification of the ICA component containing the strongest PPG signal. The second component typically contained