Abstract
Measuring objects size from an image is a challenging task and has many practical applications such as in making estimations, analyzing structure and calculating depth. In this research, we propose an end-to-end system for measuring the human foot size, which is a requirement in many systems such as in medical studies, physical studies and measurement analysis. The major challenges faced with foot size measurement were accurate extraction of foot boundary and scale variation in foot size due to its depth in the captured image. The proposed system extracts foot from the input image using color mask segmentation, which separates the foot from the noise at the background. According to our methodology, the scale variation in the foot size was normalized using a ubiquitous object as reference; in this case, a local currency coin with any face value. The reference object is captured along with the foot and afterwards separated for the depth estimation. In order to measure the foot size accurately, a linear regression model is trained and used to predict the results, and RMSE for checking the accuracy of the predicted results.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call