Automatic Surface Area and Volume Prediction on Ellipsoidal Ham using Deep Learning

Due to the emerging technologies and the rapid integration of software development, many online retailers opt to design attractive webpages in order to influence consumer perceptions of the web environment and thus leading to more sales and profits [1]. An online 2017 UPS Pulse of the Online Shopper study with more than 6,400 European respondents from France, Germany, Italy, Poland, Spain, and the UK, discovered that 84% of shoppers will shop in the physical store instead of online [2]. The major reason is that they can look around and touch and feel the products. In the in-store environment, McCabe and Nowlis [3] found that the shoppers are less likely to pick up products with geometric properties as they highly rely on the sense of vision, which is believed to provide sufficient information of the products. For some of the products with geometric properties, especially packaged products, shoppers tends to evaluate the product through visually glancing instead of reach out or touch on the product [4]. The geometry properties of an agricultural product is often indicated by its weight as it is relatively easy and quick to be measured with a digital scale. An alternative solution to obtain the weight of the product is to measure its volume as the volume is a quantity defined by mass per unit density. There are several works published to measure the volume of the objects by fusing the computer vision technique with some mathematical modeling and derivations.