GraphTEN: Graph Enhanced Texture Encoding Network

2025-03-18

AI Summary

In texture recognition, the spatial randomness of texture primitives hinders effective non-local contextual modeling. To address this, we propose a graph-enhanced multi-scale framework that jointly captures local details and global structure. First, a fully connected graph is constructed to explicitly model global pixel- or region-level dependencies. Second, a bipartite graph architecture is designed to encode cross-scale feature interactions. Third, an unordered multi-scale patch encoding module—based on a learnable codebook—is introduced to achieve scale-invariant texture representation. Our method seamlessly integrates graph neural networks (via the fully connected and bipartite graphs), CNN-based feature extraction, and multi-scale aggregation. Evaluated on five standard texture benchmarks, it consistently outperforms state-of-the-art approaches, delivering significant improvements in classification accuracy and robustness against geometric deformations and noise. This work establishes a novel paradigm for holistic texture understanding.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Graph-based Machine Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Texture recognition is a fundamental problem in computer vision and pattern recognition. Recent progress leverages feature aggregation into discriminative descriptions based on convolutional neural networks (CNNs). However, modeling non-local context relations through visual primitives remains challenging due to the variability and randomness of texture primitives in spatial distributions. In this paper, we propose a graph-enhanced texture encoding network (GraphTEN) designed to capture both local and global features of texture primitives. GraphTEN models global associations through fully connected graphs and captures cross-scale dependencies of texture primitives via bipartite graphs. Additionally, we introduce a patch encoding module that utilizes a codebook to achieve an orderless representation of texture by encoding multi-scale patch features into a unified feature space. The proposed GraphTEN achieves superior performance compared to state-of-the-art methods across five publicly available datasets.

Problem

• Modeling non-local context relations in texture recognition.
• Capturing local and global features of texture primitives.
• Achieving orderless representation of multi-scale patch features.

Innovation

• GraphTEN captures local and global texture features
• Uses fully connected and bipartite graphs for modeling
• Patch encoding module for orderless texture representation