FocalCount: Towards Class-Count Imbalance in Class-Agnostic Counting

2025-02-15

AI Summary

In category-agnostic object counting, dominant single-category images and mean squared error (MSE) loss induce severe undercounting bias for rare categories. To address this, we propose FocalCount: first, a category diversity feature estimator serves as a dynamic weighting factor to mitigate single-category dominance; second, a Focal-MSE hybrid loss function jointly optimizes MSE and binary cross-entropy (BCE), enhancing sensitivity to rare-category errors; third, multi-attribute feature fusion enables fine-grained per-category count estimation. Experiments across three counting benchmarks demonstrate substantial improvements in category-level accuracy—particularly in few-shot and zero-shot settings—while preserving overall count accuracy and significantly enhancing robustness to skewed category distributions. To our knowledge, this is the first work to jointly model category diversity and design adaptive loss functions for category-agnostic counting.

Technology Category

• Computer Vision: Object Detection & Categorization
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Large-scale security measurements

Abstract

In class-agnostic object counting, the goal is to estimate the total number of object instances in an image without distinguishing between specific categories. Existing methods often predict this count without considering class-specific outputs, leading to inaccuracies when such outputs are required. These inaccuracies stem from two key challenges: 1) the prevalence of single-category images in datasets, which leads models to generalize specific categories as representative of all objects, and 2) the use of mean squared error loss during training, which applies uniform penalization. This uniform penalty disregards errors in less frequent categories, particularly when these errors contribute minimally to the overall loss. To address these issues, we propose {FocalCount}, a novel approach that leverages diverse feature attributes to estimate the number of object categories in an image. This estimate serves as a weighted factor to correct class-count imbalances. Additionally, we introduce {Focal-MSE}, a new loss function that integrates binary cross-entropy to generate stronger error gradients, enhancing the model's sensitivity to errors in underrepresented categories. Our approach significantly improves the model's ability to distinguish between specific classes and general counts, demonstrating superior performance and scalability in both few-shot and zero-shot scenarios across three object counting datasets. The code will be released soon.

Problem

• Class-agnostic object counting
• Class-count imbalance
• Underrepresented categories inaccuracy

Innovation

• Leverages diverse feature attributes
• Introduces Focal-MSE loss function
• Corrects class-count imbalances effectively