In this paper, we propose an extension to our previous work on food portion size estimation using a single image and a multi-view volume estimation method. The single-view technique estimates food volume by using prior information (segmentation and food labels) generated from food identification methods we described earlier. For multi-view volume estimation, we use"Shape from Silhouettes"to estimate the food portion size. The experimental results of our volume estimation methods demonstrate our results with respect to accuracy and reliability.
Keywords: 3D reconstruction; dietary assessment; mobile application; portion estimation; pose estimation.