Noncontact detection refers to using external energy to detect human physiological information changes without contacting the human body. Because noncontact detection devices are exceedingly expensive and unavailable, using them on average families or for personal use is inappropriate. This study designed an image-based noncontact heartbeat measurement method for determining heartbeats by using varying light absorption on the human face. By performing image processing and independent components analysis, hidden information in the images were identified and concentrated before conducting frequency domain transform and heartbeat estimation processes, thereby achieving noncontact automatic measurement.