GCA-SUNet: A Gated Context-Aware Swin-UNet for Exemplar-Free Counting

2024-09-18

AI Summary

This paper addresses the exemplar-free object counting task—estimating object counts without category annotations or example images. We propose an end-to-end density map regression framework. Methodologically, we introduce (1) a novel gated context-aware modulation module that jointly models intra-object self-similarity attention; (2) collaborative integration of gating mechanisms into the Swin Transformer encoder, bottleneck layer, and decoder to dynamically suppress background interference; and (3) a self-similarity-guided feature enhancement strategy to improve cross-scene generalization. Evaluated on real-world benchmarks including FSC-147 and CARPK, our approach achieves significant improvements over existing state-of-the-art methods, delivering both higher accuracy and stronger generalization. The framework establishes a new paradigm for open-environment object counting, eliminating reliance on class labels or exemplar images while maintaining robustness across diverse scenes.

Technology Category

• Computer Vision: Object Detection & Categorization
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Exemplar-Free Counting aims to count objects of interest without intensive annotations of objects or exemplars. To achieve this, we propose a Gated Context-Aware Swin-UNet (GCA-SUNet) to directly map an input image to the density map of countable objects. Specifically, a set of Swin transformers form an encoder to derive a robust feature representation, and a Gated Context-Aware Modulation block is designed to suppress irrelevant objects or background through a gate mechanism and exploit the attentive support of objects of interest through a self-similarity matrix. The gate strategy is also incorporated into the bottleneck network and the decoder of the Swin-UNet to highlight the features most relevant to objects of interest. By explicitly exploiting the attentive support among countable objects and eliminating irrelevant features through the gate mechanisms, the proposed GCA-SUNet focuses on and counts objects of interest without relying on predefined categories or exemplars. Experimental results on the real-world datasets such as FSC-147 and CARPK demonstrate that GCA-SUNet significantly and consistently outperforms state-of-the-art methods. The code is available at https://github.com/Amordia/GCA-SUNet.

Problem

• Count objects without exemplars or annotations
• Map input images to density maps directly
• Suppress irrelevant features using gated mechanisms

Innovation

• Uses Swin transformers for robust feature encoding
• Implements Gated Context-Aware Modulation block
• Incorporates gate mechanism in bottleneck and decoder