Barak Hurwitz

Automatic Body Part Measurement of Dressed Humans Using Single RGB-D Camera

In the near future online shopping might be the main channel to purchase clothes. However, it is often problematic to buy clothes without trying them first. We present a novel methodology to measure human body parts that allows users to wear loose clothes during this process. This allows anyone at home or on the go to measure themselves and use the results with their 3D model to comfortably shop online. Reducing the risk of buying clothes that will not fit and making the online shopping experience more satisfying. We first scan the user on a single pose that allows better visualization of the waist and then measure four parts: chest, waist and low hip girths and height. Then, we use an API to compute the rest of the measurements. Clothes sizes differ in 4 to 5 cm and our results indicate an offset error of less than 4 cm. Hence, meeting our target immersive shopping use cases.