Learning to assess visual aesthetics of food images
- PDF / 6,077,623 Bytes
- 14 Pages / 612 x 808 pts Page_size
- 45 Downloads / 173 Views
Learning to assess visual aesthetics of food images Kekai Sheng1,2, Weiming Dong2 ( ), Haibin Huang3, Menglei Chai4, Yong Zhang5, Chongyang Ma3, and Bao-Gang Hu2 c The Author(s) 2020.
Abstract Distinguishing aesthetically pleasing food photos from others is an important visual analysis task for social media and ranking systems related to food. Nevertheless, aesthetic assessment of food images remains a challenging and relatively unexplored task, largely due to the lack of related food image datasets and practical knowledge. Thus, we present the Gourmet Photography Dataset (GPD), the first largescale dataset for aesthetic assessment of food photos. It contains 24,000 images with corresponding binary aesthetic labels, covering a large variety of foods and scenes. We also provide a non-stationary regularization method to combat over-fitting and enhance the ability of tuned models to generalize. Quantitative results from extensive experiments, including a generalization ability test, verify that neural networks trained on the GPD achieve comparable performance to human experts on the task of aesthetic assessment. We reveal several valuable findings to support further research and applications related to visual aesthetic analysis of food images. To encourage further research, we have made the GPD publicly available at https://github. com/Openning07/GPA.
1
Introduction
Food is one of the most fundamental entities in our daily life. A great food photograph can convey feelings of warmth, awaken fond memories, conjure up fantasies, or just simply make you hungry [1]. It can also draw crowds flocking to a new restaurant or boost the sales of a food magazine. Thus, the ability to assess the aesthetic quality of food images plays an important role in various applications, such as food photo recommendation (see Fig. 1(a)), food photography assistance, and enhancement (Fig. 1(b)). Human beings can easily gauge the visual aesthetics of food photos. However, it remains challenging for artificial intelligent agents to do so. During the past two decades, many researchers have considered various related fields, such as image aesthetic assessment [2–4] and food image analysis [5–7]. Some have already explored aesthetic assessment of food
Keywords image aesthetic assessment; food image analysis; dataset; regularization 1 Youtu Lab, Tencent, Shanghai 200233, China. E-mail: [email protected]. 2 NLPR, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China. E-mail: W. Dong, [email protected] ( ); B.-G. Hu, [email protected]. 3 Kuaishou Technology, Beijing 100085, China. E-mail: H. Huang, [email protected]; C. Ma, [email protected]. 4 Snap Inc., Santa Monica, 90405, USA. E-mail: [email protected]. 5 AI Lab, Tencent Inc., Shenzhen 518000, China. E-mail: [email protected]. Manuscript received: 2020-06-09; accepted: 2020-08-25
Fig. 1 (a) When you browse photos of strawberry cake, it would be nice if photos are sorted by visual aesthetics, e.g., aesthetically negative (top) and positive ones (b
Data Loading...
