Use Classifier as Generator
This addresses the challenge of image generation without retraining, though it is incremental with limited quality.
The paper tackles the problem of image generation by proposing a method to directly use a normally trained classifier for this task, achieving recognizable but limited-quality results on MNIST.
Image recognition/classification is a widely studied problem, but its reverse problem, image generation, has drawn much less attention until recently. But the vast majority of current methods for image generation require training/retraining a classifier and/or a generator with certain constraints, which can be hard to achieve. In this paper, we propose a simple approach to directly use a normally trained classifier to generate images. We evaluate our method on MNIST and show that it produces recognizable results for human eyes with limited quality with experiments.