Learning To Count Everything

TL;DR

This paper introduces a novel few-shot regression approach for counting objects of any category in images, using minimal annotations and an adaptation strategy, supported by a new dataset of 147 categories.

Contribution

The paper proposes a new few-shot counting method that generalizes to any category and introduces a comprehensive dataset for this task.

Findings

Outperforms existing object detectors and few-shot counting methods

Effective adaptation to new categories with few exemplars

Provides a new dataset for diverse object counting

Abstract

Existing works on visual counting primarily focus on one specific category at a time, such as people, animals, and cells. In this paper, we are interested in counting everything, that is to count objects from any category given only a few annotated instances from that category. To this end, we pose counting as a few-shot regression task. To tackle this task, we present a novel method that takes a query image together with a few exemplar objects from the query image and predicts a density map for the presence of all objects of interest in the query image. We also present a novel adaptation strategy to adapt our network to any novel visual category at test time, using only a few exemplar objects from the novel category. We also introduce a dataset of 147 object categories containing over 6000 images that are suitable for the few-shot counting task. The images are annotated with two types of annotation, dots and bounding boxes, and they can be used for developing few-shot counting models. Experiments on this dataset shows that our method outperforms several state-of-the-art object detectors and few-shot counting approaches. Our code and dataset can be found at https://github.com/cvlab-stonybrook/LearningToCountEverything.