TREND: Truncated Generalized Normal Density Estimation of Inception Embeddings for GAN Evaluation
This addresses the need for more accurate and robust evaluation metrics for GANs in image generation, though it is incremental as it builds upon the widely used Frechét Inception Distance.
The paper tackles the problem of unreliable GAN evaluation by proposing TREND, a metric that uses truncated generalized normal distribution for density estimation of Inception embeddings, significantly reducing errors and improving robustness against sample size variation.
Evaluating image generation models such as generative adversarial networks (GANs) is a challenging problem. A common approach is to compare the distributions of the set of ground truth images and the set of generated test images. The Frechét Inception distance is one of the most widely used metrics for evaluation of GANs, which assumes that the features from a trained Inception model for a set of images follow a normal distribution. In this paper, we argue that this is an over-simplified assumption, which may lead to unreliable evaluation results, and more accurate density estimation can be achieved using a truncated generalized normal distribution. Based on this, we propose a novel metric for accurate evaluation of GANs, named TREND (TRuncated gEneralized Normal Density estimation of inception embeddings). We demonstrate that our approach significantly reduces errors of density estimation, which consequently eliminates the risk of faulty evaluation results. Furthermore, we show that the proposed metric significantly improves robustness of evaluation results against variation of the number of image samples.