Machine Learning in Image Processing and Computer Vision

Dear Colleagues,

Recent advances in machine learning have dramatically transformed the capabilities of image processing and computer vision, opening new possibilities for detailed visual understanding. One of the most challenging and intellectually demanding areas in this field is fine-grained object recognition, the task of distinguishing between visually similar categories characterized by subtle differences and high intra-class variability. This issue has become critical in applications such as retail product identification, animal breed classification, or precision inspection in manufacturing, where recognizing visual nuances is essential.

This Special Issue aims to showcase recent developments that address these challenges using robust learning architectures, semantic modeling, and efficient recognition techniques with high resolutions and low semantic granularity.

We especially encourage submissions that explore a broad range of modern methodologies, including, but not limited to, the following:

• Prototype-based and metric learning methods which enhance interpretability and class separability by modeling data around representative exemplars;
• Proxy-based learning allowing for scalable training and inference in classification tasks with large or imbalanced category sets;
• Domain adaptation and domain generalization, essential for transferring models across visually diverse environments (e.g., synthetic-to-real, controlled-to-wild conditions);
• Few-shot and meta-learning strategies enabling models to generalize from limited labeled examples or rapidly adapt to novel classes;
• Contrastive learning and self-supervised pretraining, which improve feature quality and sample efficiency, especially in data-scarce settings;
• Knowledge distillation to compress and transfer information from large high-capacity models into lightweight deployable ones;
• Attention mechanisms, transformers, and hierarchical learning, which help capture the multi-scale and contextual dependencies critical to fine-grained discrimination;
• Multimodal fusion (e.g., combining visual and textual cues), useful in applications like product catalog matching or e-commerce;
• Explainable AI (XAI) approaches for interpreting model decisions in high-stakes or user-facing applications.

We also welcome research on dataset creation, benchmark challenges, and performance evaluation methodologies tailored to fine-grained problems.

This Special Issue provides a forum for researchers, engineers, and practitioners to present their latest results, share new tools and datasets, and contribute to the collective advancement of fine-grained image understanding powered by modern machine learning.

Dr. Grzegorz Sarwas
Guest Editor

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 100 words) can be sent to the Editorial Office for announcement on this website.

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. Information 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 1600 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.

• fine-grained object recognition
• prototype and proxy learning
• domain adaptation in computer vision
• knowledge distillation
• few-shot and metric learning