Applications and Effect Evaluation of Generative Adversarial Networks in Semi-Supervised Learning
Jiyu Hu, Haijiang Zeng, Zhen Tian
TL;DR
This paper addresses semi-supervised image classification under limited labeled data by introducing a GAN-based framework that couples a generator, discriminator, and classifier in a collaborative training loop. The generator employs an encoder–decoder with multi-head self-attention, while the classifier uses conditional batch normalization to leverage category information; the discriminator includes a category-aware header to fuse semantic signals. Empirical results on MNIST and SVHN show the approach outperforms traditional semi-supervised GANs and CNNs in both generation quality and classification accuracy, particularly in low-label regimes (e.g., 10% labels). The work demonstrates a robust, generalizable strategy for semi-supervised learning in image classification and points toward broader applications in vision tasks such as object detection and segmentation.
Abstract
In recent years, image classification, as a core task in computer vision, relies on high-quality labelled data, which restricts the wide application of deep learning models in practical scenarios. To alleviate the problem of insufficient labelled samples, semi-supervised learning has gradually become a research hotspot. In this paper, we construct a semi-supervised image classification model based on Generative Adversarial Networks (GANs), and through the introduction of the collaborative training mechanism of generators, discriminators and classifiers, we achieve the effective use of limited labelled data and a large amount of unlabelled data, improve the quality of image generation and classification accuracy, and provide an effective solution for the task of image recognition in complex environments.