The augmented human - visual movement magnification

• Main Author: Tan, Ryan Jinn-En
• Other Authors: Cham Tat Jen
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: https://hdl.handle.net/10356/166679
• Institution: Nanyang Technological University

Over the years of mankind’s history, the advent of tools has allowed us to extend our human capabilities. Each of the 5 senses (Touch, Hearing, Smell, Taste, and most importantly Sight) have received countless enhancements through the power of human ingenuity and creativity, borrowing elements from nature, science and sometimes our imagination granting us the ability to achieve feats that continue to scale in both scope and aspirations. This report discusses utilising an existing Learning-based Video Motion Magnification (LVMM) that allows users to observe previously indiscernible movement such as breathing, pulse and tiny facial movements of other people from a video recorded using conventional camera equipment (such as a mobile phone), before feeding the output data into a separate deep-learning image classification model built using Keras to train it into discerning between a person inhaling and exhaling. The image classification software will serve as a foundation for monitoring a person’s respiratory cycle and potentially be used in conjunction with existing medical devices to further expand the groundwork for non-invasive patient care.