Relative Pixel Prediction For Autoregressive Image Generation

Wang Ling, Chris Dyer, Lei Yu, Lingpeng Kong, Dani Yogatama, Susannah Young

Keywords: Image Generation, Autoregressive

Abstract:

In natural images, transitions between adjacent pixels tend to be smooth and gradual, a fact that has long been exploited in image compression models based on predictive coding. In contrast, existing neural autoregressive image generation models predict the absolute pixel intensities at each position, which is a more challenging problem. In this paper, we propose to predict pixels relatively, by predicting new pixels relative to previously generated pixels (or pixels from the conditioning context, when available). We show that this form of prediction fare favorably to its absolute counterpart when used independently, but their coordination under an unified probabilistic model yields optimal performance, as the model learns to predict sharp transitions using the absolute predictor, while generating smooth transitions using the relative predictor.
Experiments on multiple benchmarks for unconditional image generation, image colorization, and super-resolution indicate that our presented mechanism leads to improvements in terms of likelihood compared to the absolute prediction counterparts.