Granular-ball Guided Masking: Structure-aware Data Augmentation

2025-12-24

AI Summary

Deep models suffer from overfitting in few-shot learning and under distributional shift, while existing mask-based data augmentation methods lack structural awareness. Method: This paper proposes a structure-aware granular-ball-guided data augmentation method. It introduces Granular Ball Computing (GBC) into data augmentation for the first time, employing a coarse-to-fine hierarchical masking strategy that adaptively preserves semantically rich and structurally critical regions while suppressing redundant information. The method constitutes a hierarchical, model-agnostic, plug-and-play framework compatible with both CNNs and Vision Transformers (ViTs). Contribution/Results: Evaluated on multiple benchmarks, it significantly improves classification accuracy and enhances masked image reconstruction quality. Crucially, it achieves adaptive mask generation jointly guided by semantic and structural cues—constituting the core innovation—thereby comprehensively strengthening model generalization and robustness under limited-data and out-of-distribution scenarios.

Technology Category

• Machine Learning: Structured Learning
• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Learning to Search

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Deep learning models have achieved remarkable success in computer vision, but they still rely heavily on large-scale labeled data and tend to overfit when data are limited or distributions shift. Data augmentation, particularly mask-based information dropping, can enhance robustness by forcing models to explore complementary cues; however, existing approaches often lack structural awareness and may discard essential semantics. We propose Granular-ball Guided Masking (GBGM), a structure-aware augmentation strategy guided by Granular-ball Computing (GBC). GBGM adaptively preserves semantically rich, structurally important regions while suppressing redundant areas through a coarse-to-fine hierarchical masking process, producing augmentations that are both representative and discriminative. Extensive experiments on multiple benchmarks demonstrate consistent improvements in classification accuracy and masked image reconstruction, confirming the effectiveness and broad applicability of the proposed method. Simple and model-agnostic, it integrates seamlessly into CNNs and Vision Transformers and provides a new paradigm for structure-aware data augmentation.

Problem

• Enhances robustness of deep learning models with limited data
• Preserves semantically rich regions while suppressing redundant areas
• Improves classification accuracy and masked image reconstruction

Innovation

• Granular-ball Computing guides structure-aware masking
• Coarse-to-fine hierarchical masking preserves semantic regions
• Model-agnostic augmentation integrates into CNNs and Transformers