IEEE Transactions on Pattern Analysis and Machine Intelligence
Call for Papers
Special Issue on Fine-Grained Visual Categorization
Aims and Scope
With the techniques for standard supervised image classification becoming increasingly practical, fine-
grained image categorization where images are classified into subordinate categories, have recently
attracted a lot of attention and become an important task in computer vision. Examples of fine-grained
visual categorization include but are not limited to recognizing, e.g., detailed animal species, specific sub-
groups of plants, and car makers and models. Compared with the general-purpose image categorization
tasks, such as the ImageNet Challenge of 1,000 general categories, fine-grained categorization pays
attention to subtle details that are not easily captured using the off-the-shelf image classifiers -- a
promising direction in visual perception and image understanding beyond generic labels. In addition, the
absence of sufficient training data with the presence of a large number of fine-grained categories, e.g.,
about 10K species for birds and over 250K species for flowers, makes the problem of fine-grained visual
categorization particularly interesting and challenging.
Applying deep neural networks, which are proposed for general-purpose image classification, to fine-
grained visual categorization has led to notable performance improvement, but the fine-grained
categorization problem cannot be solved merely by training modern deep convolutional neural networks.
In the past, results for fine-grained image categorization have been mostly attained using classifiers with
strong supervision, where detailed labels such as body parts, attributes, and viewpoints are manually
annotated and used in training. Many questions generally arise when the fine-grained categorization task
is made more general and broad: How do we alleviate the burden of having fine-grained manual
annotations? How can top-down information and domain knowledge be included? How can we make the
best use of web data and online resources like Mechanical Turk?
Topics and Guidelines
This special issue targets researchers and practitioners from both industry and academia to provide a forum
in which to publish recent state-of-the-art achievements in the fine-grained image recognition area. Topics
of interest include, but are not limited to:
• Fine-grained image recognition and categorization
• Fine-grained vehicle categorization and verification
• Visual logo detection, categorization and verification
• Person re-identification
• Vehicle re-identification
• Fashion image recognition, search and attribute prediction
• Fine-grained food recognition and ingredient analysis
• Transfer-learning from categories to subcategories
• Part-based models for fine-grained categorization
• Attribute-based models for fine-grained feature learning
• Ontology-based fine-grained visual categorization
• Fine-grained categorization with human in the loop
• Fine-grained categorization in the wild by domain adaptation
• Multi-modal data for fine-grained categorization
• Novel benchmark data
• Novel annotation, crowdsourcing approaches and tools for fine-grained data labeling
Before submitting your manuscript, please ensure you have carefully read the Instructions for Authors for
IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI). The complete manuscript
should be submitted through T-PAMI’s submission system (https://mc.manuscriptcentral.com/tpami-cs).
To ensure that you submit to the correct special issue, please select the appropriate section in the drop-
down menu upon submission. In your cover letter, please also clearly mention the title of the SI.
Important Dates
• Paper submission due: August 31, 2018
• First notification: November 30, 2018
• Revision: January 31, 2019
• Final decision: Aril 15, 2019
• Publication date: July 2019 (Tentative)
Guest Editors
Jingdong Wang, Microsoft Research, Beijing, China
Zhuowen Tu, University of California, San Diego, US
Jianlong Fu, Microsoft Research, Beijing, China
Nicu Sebe, University of Trento, Italy
Serge Belongie, Cornell University & Cornell Tech, US
Contact email: jianf@microsoft.com, jingdw@microsoft.com
