Efficient and Accurate Quantized Image Super-Resolution on Mobile NPUs, Mobile AI & AIM 2022 challenge: Report

TL;DR

This paper reports on a challenge to develop efficient, quantized image super-resolution models suitable for real-time deployment on mobile NPUs, achieving high-quality 3X upscaling at 60 FPS on a specific edge NPU.

Contribution

The paper introduces a set of optimized quantized super-resolution models designed for mobile NPUs, demonstrating real-time performance and high-quality image upscaling.

Findings

Models achieve up to 60 FPS on Full HD images

All solutions are compatible with Synaptics VS680 NPU

High-quality 3X image upscaling on mobile devices

Abstract

Image super-resolution is a common task on mobile and IoT devices, where one often needs to upscale and enhance low-resolution images and video frames. While numerous solutions have been proposed for this problem in the past, they are usually not compatible with low-power mobile NPUs having many computational and memory constraints. In this Mobile AI challenge, we address this problem and propose the participants to design an efficient quantized image super-resolution solution that can demonstrate a real-time performance on mobile NPUs. The participants were provided with the DIV2K dataset and trained INT8 models to do a high-quality 3X image upscaling. The runtime of all models was evaluated on the Synaptics VS680 Smart Home board with a dedicated edge NPU capable of accelerating quantized neural networks. All proposed solutions are fully compatible with the above NPU, demonstrating an up to 60 FPS rate when reconstructing Full HD resolution images. A detailed description of all models developed in the challenge is provided in this paper.