NaturalInversion: Data-Free Image Synthesis Improving Real-World Consistency
This addresses the challenge of data-free image synthesis for tasks like model compression, though it appears incremental as it builds on existing model inversion methods.
The paper tackles the problem of synthesizing images that match the original data distribution without using real data, and results show that their method produces images more consistent with the original distribution and outperforms prior works in applications like knowledge distillation and pruning.
We introduce NaturalInversion, a novel model inversion-based method to synthesize images that agrees well with the original data distribution without using real data. In NaturalInversion, we propose: (1) a Feature Transfer Pyramid which uses enhanced image prior of the original data by combining the multi-scale feature maps extracted from the pre-trained classifier, (2) a one-to-one approach generative model where only one batch of images are synthesized by one generator to bring the non-linearity to optimization and to ease the overall optimizing process, (3) learnable Adaptive Channel Scaling parameters which are end-to-end trained to scale the output image channel to utilize the original image prior further. With our NaturalInversion, we synthesize images from classifiers trained on CIFAR-10/100 and show that our images are more consistent with original data distribution than prior works by visualization and additional analysis. Furthermore, our synthesized images outperform prior works on various applications such as knowledge distillation and pruning, demonstrating the effectiveness of our proposed method.