Learning to Upsample by Learning to Sample

TL;DR

DySample is a lightweight, resource-efficient dynamic upsampler that outperforms existing methods across multiple dense prediction tasks by reformulating upsampling as a point sampling problem, avoiding complex dynamic convolutions.

Contribution

The paper introduces DySample, a novel upsampling method that bypasses dynamic convolution, reducing computational load and broadening application scenarios while achieving superior performance.

Findings

DySample requires no custom CUDA code.

DySample has fewer parameters, FLOPs, and memory usage.

DySample outperforms other upsamplers in five dense prediction tasks.

Abstract

We present DySample, an ultra-lightweight and effective dynamic upsampler. While impressive performance gains have been witnessed from recent kernel-based dynamic upsamplers such as CARAFE, FADE, and SAPA, they introduce much workload, mostly due to the time-consuming dynamic convolution and the additional sub-network used to generate dynamic kernels. Further, the need for high-res feature guidance of FADE and SAPA somehow limits their application scenarios. To address these concerns, we bypass dynamic convolution and formulate upsampling from the perspective of point sampling, which is more resource-efficient and can be easily implemented with the standard built-in function in PyTorch. We first showcase a naive design, and then demonstrate how to strengthen its upsampling behavior step by step towards our new upsampler, DySample. Compared with former kernel-based dynamic upsamplers, DySample requires no customized CUDA package and has much fewer parameters, FLOPs, GPU memory, and latency. Besides the light-weight characteristics, DySample outperforms other upsamplers across five dense prediction tasks, including semantic segmentation, object detection, instance segmentation, panoptic segmentation, and monocular depth estimation. Code is available at https://github.com/tiny-smart/dysample.