Computer Vision for Food Data Analysis: Methods, Challenges, and Applications

Dear Colleagues,

Computer vision has achieved remarkable progress in various downstream tasks, demonstrating promising performance and application potential in real-world scenarios. However, food image analysis represents a domain where current computer vision algorithms face significant limitations, revealing a substantial performance gap between general computer vision tasks and food-specific applications. This gap stems from several unique characteristics of food data, such as (1) intra-class variations (the same dish can appear drastically different), (2) inter-class similarities (different dishes may look very similar), (3) complex compositional structures (multiple ingredients with various preparation methods), and (4) significant cultural diversities. These challenges are further compounded by the nature of real-world food data available on the internet, which is often ambiguous, non-annotated, noisy, watermarked, and of varying quality.

The importance of addressing these computer vision challenges has grown rapidly with the widespread adoption of smartphones and social media platforms, which continuously generate massive amounts of food-related visual data. While food serves as a fundamental element of human society, the technical challenges it presents make it an important domain for advancing computer vision capabilities. The successful development of robust computer vision systems for food analysis could enable breakthrough applications across multiple domains, from automated dietary monitoring and nutritional assessment systems to cross-cultural food recognition systems. Furthermore, with an estimated 200,000 basic dishes globally, food recognition presents fundamental computer vision research questions regarding classification scalability with massive numbers of fine-grained categories, robustness to noisy and limited labels, and the development of more sophisticated visual understanding models.

Recent advances in computer vision, particularly in generative AI, large language models, and multi-modal learning, present promising new directions for addressing these challenges. This Special Issue aims to explore innovative approaches in advancing our ability to analyze and understand food through visual data. We seek contributions that advance both fundamental computer vision techniques and practical applications in food-related tasks. Submissions may include, but are not limited to, the following topics:

• Food image/video generation and generative AI;
• Food video analysis and action recognition;
• Food 3D model reconstruction;
• Food portion/nutrition value estimation;
• Food manipulation understanding;
• Food image quality analysis/inspection;
• Eating and cooking action recognition;
• Multi-modal food data analysis;
• Food ontologies and LLM-based models for food data analysis;
• Visual question answering for food;
• Food data analysis and uncertainty modeling;
• Learning with noisy food labels;
• Continual, self-supervised, semi-supervised, and unsupervised learning for food;
• Food classification/detection/segmentation in 2D/3D.

Dr. Jiangpeng He
Dr. Luotao Lin
Dr. Bhalaji Nagarajan
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 250 words) can be sent to the Editorial Office for assessment.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Journal of Imaging is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 1800 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• computer vision
• food image analysis
• generative AI
• deep learning
• multi-modal analysis
• action recognition
• 3D reconstruction
• visual understanding
• food recognition
• machine learning