This study explores a non-contact method for estimating stress using facial images captured by a smartphone ‘s built-in camera or webcam. By analyzing RGB values, subtle changes in facial blood flow are detected to extract pulse-like signals and calculate RR intervals. Frequency analysis is then performed to determine the LF/HF ratio, a key stress indicator. Higher LF/HF ratios typically indicate greater stress levels. The goal is to achieve accuracy comparable to traditional PPG-based heart rate variability analysis while maintaining a non-contact approach. Advanced data processing and refined frequency analysis are applied to enhance reliability. Experiments will evaluate performance in real-world conditions, considering factors like measurement environments and image acquisition settings. This method aims to provide a practical, non-invasive, and efficient solution for remote stress assessment.