A Re-node Self-training Approach for Deep Graph-based Semi-supervised Classification on Multi-view Image Data

TL;DR

The paper tackles semi-supervised learning on multi-view image data that lack explicit graph structures by introducing RSGSLM, a three-module framework that learns per-view graphs, fuses them into a single adaptive graph, and uses a GCN with a dynamically weighted pseudo-label loss. It integrates ReNode-based topological balancing and manifold regularization, enabling effective learning with limited labels and unlabeled data. Key contributions include a shallow, semi-supervised graph construction per view, adaptive view fusion with smoothness-based weighting, and a temporally scaled pseudo-label loss that eliminates fixed thresholds, all within a single GCN model. Empirical results on six benchmark datasets show state-of-the-art performance with improved efficiency, highlighting practical impact for multi-view visual tasks under scarce supervision, and suggesting pathways for future work in inductive learning, transfer learning, and attention mechanisms.

Abstract

Recently, graph-based semi-supervised learning and pseudo-labeling have gained attention due to their effectiveness in reducing the need for extensive data annotations. Pseudo-labeling uses predictions from unlabeled data to improve model training, while graph-based methods are characterized by processing data represented as graphs. However, the lack of clear graph structures in images combined with the complexity of multi-view data limits the efficiency of traditional and existing techniques. Moreover, the integration of graph structures in multi-view data is still a challenge. In this paper, we propose Re-node Self-taught Graph-based Semi-supervised Learning for Multi-view Data (RSGSLM). Our method addresses these challenges by (i) combining linear feature transformation and multi-view graph fusion within a Graph Convolutional Network (GCN) framework, (ii) dynamically incorporating pseudo-labels into the GCN loss function to improve classification in multi-view data, and (iii) correcting topological imbalances by adjusting the weights of labeled samples near class boundaries. Additionally, (iv) we introduce an unsupervised smoothing loss applicable to all samples. This combination optimizes performance while maintaining computational efficiency. Experimental results on multi-view benchmark image datasets demonstrate that RSGSLM surpasses existing semi-supervised learning approaches in multi-view contexts.

Figures (5)

• Figure 1: The framework of our proposed RSGSLM. The model comprises three distinct modules: the first module is dedicated to graph learning and feature transformation, the second module handles topological imbalance adjustment, and the third module performs graph convolution.
• Figure 2: Sensitivity to parameters $\lambda_1$ and $\lambda_2$ on the Scene and Youtube datasets.
• Figure 3: Sensitivity to parameter $w_{max}-w_{min}$ on the Scene dataset.
• Figure 4: The t-SNE visualization of $\hbox{\bf X}_\ast$ and $\hbox{\bf Z}$ in RSGSLM on four datasets.
• Figure 5: The t-SNE visualization of $\hbox{\bf X}_\ast$ and output of three models on ALOI.