The performance of learning-based methods is adversely affected by imbalance problems at various levels, including the input, intermediate or mid-level stages of the processing or the objectives to be optimized in a multi-task setting. Currently, researchers tend to address these challenges in their particular contexts with problem-specific solutions and with limited awareness of the solutions proposed for similar challenges in other computer vision problems.

Imbalance problems pertain to almost all computer vision problems and therefore, the workshop is highly relevant and interesting for a broad community. Moreover, since learning-based approaches dominate the current methodologies and systems, a workshop that targets general issues in such learning approaches is highly beneficial. In addition, since the availability of data does not solve imbalance problems, research interest in addressing imbalance problems is going to increase more.

Our recent review paper on imbalance problems in object detection ("Imbalance problems in object detection: a review", accepted for publication at IEEE TPAMI) cites over 200 papers which were written by 655 unique authors. We interpret these numbers (which are specific to object detection) as strong indicators of interest in imbalance problems.

In an effort to quantify the interest in general, we searched for publications (using www.dimensions.ai) that included the phrase "imbalance problems in computer vision'' in the text (note that including this phrase does not necessarily mean that the paper is on this subject; however, it gives an indication of the potential span of the interest). The number of publications and citations with this phrase is plotted in the following figure, where we see a significantly increasing interest on research on this subject. These publications amount to more than 2000 unique authors (number of authors with more than 1 papers with the matching phrase), which indicates the vast community that is interested in this problem.