Height-and-Weight-Estimation-of-an-Individual-from-Virtual-Visuals

The rapid growth of technology has made people lead digital dynamic lives, where technology dominates normal human life. With the advancements in technology, there is a constant update to new methods, and techniques replacing the outdated existing methodologies. Adapting to the improvements in technology, this paper proposes a novel approach to estimate the height and weight of a person. Despite using measuring instruments, the orthodox style of knowing these attributes of a person, these values can be computed with the use of a person’s visuals in real-time or virtual images. After leveraging some considerable amount of attention to the human face, the information that a face can exhibit motivated us to predict an individual’s height, and weight by using facial traits. Using this information, and available applications, a novel algorithm has been developed to estimate the height, and weight of an individual with the help of either a photograph or video consisting full-length image. In different to other previous models, our model predicts the height and weight of an individual without any external reference object in the video frame or the photograph. And when this model is tested with 50 full-length images acceptable results were seen, with the mean absolute height difference of 6.9 cms, and 3.2 kgs of the weight difference between predicted and original values. OpenCV, Python, Mediapipe, and Machine learning algorithms are the technologies used to implement this gender-based height and weight prediction system.