Debiasing Kernel-Based Generative Models
This work addresses image quality issues in generative models for computer vision applications, but it is incremental as it builds on existing kernel and debiasing techniques.
The authors tackled the problem of blurry image generation in kernel-based generative models by proposing a two-stage framework that uses kernel density estimation and a debiasing algorithm, achieving image quality on CIFAR10 comparable to state-of-the-art models like diffusion and GANs, with competitive results on CelebA and LSUN datasets.
We propose a novel two-stage framework of generative models named Debiasing Kernel-Based Generative Models (DKGM) with the insights from kernel density estimation (KDE) and stochastic approximation. In the first stage of DKGM, we employ KDE to bypass the obstacles in estimating the density of data without losing too much image quality. One characteristic of KDE is oversmoothing, which makes the generated image blurry. Therefore, in the second stage, we formulate the process of reducing the blurriness of images as a statistical debiasing problem and develop a novel iterative algorithm to improve image quality, which is inspired by the stochastic approximation. Extensive experiments illustrate that the image quality of DKGM on CIFAR10 is comparable to state-of-the-art models such as diffusion models and GAN models. The performance of DKGM on CelebA 128x128 and LSUN (Church) 128x128 is also competitive. We conduct extra experiments to exploit how the bandwidth in KDE affects the sample diversity and debiasing effect of DKGM. The connections between DKGM and score-based models are also discussed.