Personalized Food Printing for Portrait Images

Computers and Graphics, 2018, 70: 188-197.

Haiming Zhao, Jufeng Wang, Xiaoyu Ren, Jingyuan Li, Yong-Liang Yang, and Xiaogang Jin

The overview of our framework. Given an input image (g), our framework applies image abstraction (h) and path optimization (i), to generate personalized pattern that can be directly fabricated (j). Sketch synthesis (aCb) and face enhancement (cCd) can be further employed to improve the printing quality of portrait images.


The recent development of 3D printing techniques enables novel applications in customized food fabrication. Based on a tailor-made 3D food printer, we present a novel personalized food printing framework driven by portrait images. Unlike common 3D printers equipped with materials such as ABS, Nylon and SLA, our printer utilizes edible materials such as maltose, chocolate syrup, jam to print customized patterns. Our framework automatically converts an arbitrary input image into an optimized printable path to facilitate food printing, while preserving the prominent features of the image. This is achieved based on two key stages. First, we apply image abstraction techniques to extract salient image features. Robust face detection and sketch synthesis are optionally involved to enhance face features for portrait images. Second, we present a novel path optimization algorithm to generate printing path for efficient and feature-preserving food printing. We demonstrate the efficiency and efficacy of our framework using a variety of images and also a comparison with non-optimized results.