Projects Last Projects

Early detection of respiratory distress is critical for timely medical intervention, particularly in
young children. This study presents a non-invasive, video-based approach for assessing respiratory
effort by analyzing depth variations of the torso segmentation. Using a combination of depth
estimation models, segmentation models, and signal processing techniques, we extract chest and
abdomen respiratory rates (Breaths per minute) & phase angle between the signals. The data
undergoes noise reduction, Gaussian-based depth calculations, and signal processing to enhance
signal clarity. This framework demonstrates the potential for automated, contact-free respiratory
assessment, offering a promising tool for early detection and continuous monitoring in clinical and
home environments.

Please, see project report.

Please, see final presentation.