Fast and Accurate Quantized Camera Scene Detection on Smartphones, Mobile AI 2021 Challenge: Report

TL;DR

This paper reports on a challenge to develop quantized deep learning models for real-time camera scene detection on smartphones, achieving high accuracy and speed on mobile platforms.

Contribution

It introduces the first Mobile AI challenge for camera scene detection, providing a large dataset and benchmarking models on real smartphone hardware.

Findings

Models achieve 100-200 FPS on recent smartphones

Top-3 accuracy exceeds 98% on the dataset

Solutions are compatible with major mobile AI accelerators

Abstract

Camera scene detection is among the most popular computer vision problem on smartphones. While many custom solutions were developed for this task by phone vendors, none of the designed models were available publicly up until now. To address this problem, we introduce the first Mobile AI challenge, where the target is to develop quantized deep learning-based camera scene classification solutions that can demonstrate a real-time performance on smartphones and IoT platforms. For this, the participants were provided with a large-scale CamSDD dataset consisting of more than 11K images belonging to the 30 most important scene categories. The runtime of all models was evaluated on the popular Apple Bionic A11 platform that can be found in many iOS devices. The proposed solutions are fully compatible with all major mobile AI accelerators and can demonstrate more than 100-200 FPS on the majority of recent smartphone platforms while achieving a top-3 accuracy of more than 98%. A detailed description of all models developed in the challenge is provided in this paper.