Variance Constrained Distribution Alignment in Few-shot Models

Few-shot image generation aims to learn class-conditional generative models from limited data. Existing approaches often suffer from intra-class distribution drift and poor generalization due to unstable estimation of class-level representations. To address these challenges, we propose a method that models class-level latent distributions for flexible and efficient few shot synthesis. Specifically, each input is represented by a learnable conditional latent distribution. Metric-based statistical modeling effectively disentangles latent variables, contracts intra-class variance, and enlarges inter-class margins while enforcing cross-task distributional alignment. We further provide a variance-based generalization analysis, showing that controlling class-conditional variance tightens generalization bounds under small-sample regimes. Experiments on benchmark datasets demonstrate that our method surpasses prior works in visual quality and diversity, highlighting the benefit of statistical alignment for robust few-shot generative modeling.