MAI: Mobile AI workshop and challenges 2022
In conjunction with CVPR 2022

https://ai-benchmark.com/workshops/mai/2022/
Contact: radu.timofte [at] uni-wuerzburg.de

Scope

Over the past years, mobile AI-based applications are becoming more and more ubiquitous. Various deep learning models can now be found on any mobile device starting from smartphones running portrait segmentation, image enhancement, face recognition and natural language processing models, to IoT platforms performing real-time image classification or smart-TV boards coming with sophisticated image super-resolution algorithms. The performance of mobile NPUs and DSPs is also increasing dramatically, making it possible to run complex deep learning models and to achieve fast runtime in the majority of tasks.

While many research papers targeted at efficient deep learning models have been proposed recently, the evaluation of the obtained solutions is usually happening on desktop CPUs and GPUs, making it nearly impossible to estimate the actual inference time and memory consumption on real mobile hardware. To address this problem, we introduce the first Mobile AI Workshop, where all solutions and deep learning models will be evaluated on the actual mobile AI accelerators.

The workshop will consist of the three main parts, including: 1) a detailed overview of deep learning inference on mobile platforms, 2) workshop challenges where the participants can get the actual hands-on experience while solving several computer vision tasks and evaluating their solutions on mobile devices, and 3) presentations from mobile SoC vendors covering several important aspects of mobile AI inference.

To ensure that the participants get the most recent and actual information on mobile-related AI tasks, the workshop is designed in collaboration with several major mobile SoC vendors, including Qualcomm, Samsung, Huawei, MediaTek, and Synaptics.

This workshop also builds upon the success of the previous computer vision competitions and is organized by people associated with the NTIRE (CVPR 2017, 2018, 2019, 2020 and 2021), CLIC (2018, 2019, 2020, 2021), PIRM (2018) and AIM (2019, 2020, 2021) workshops.

Topics

Papers addressing the topics covering efficient deep learning for mobile devices, mobile-based vision / natural language processing / performance evaluation are invited. The topics include, but are not limited to:

● Efficient deep learning models for mobile devices
● Artifacts removal from mobile photos/videos
● General smartphone photo/video enhancement
● RAW camera image/video processing
● Deep learning applications for mobile camera ISPs
● Image/video super-resolution on low-power hardware
● Portrait segmentation / bokeh effect rendering
● Depth estimation w/o multiple cameras
● Perceptual image manipulation on mobile devices
● Activity recognition using smartphone sensors
● Image/sensor based identity recognition
● Fast image classification / object detection algorithms
● NLP models optimized for mobile inference
● Real-time semantic segmentation
● Low-power machine learning inference
● Machine learning and deep learning frameworks for mobile devices
● AI performance evaluation / benchmarking of mobile and IoT hardware
● Studies and applications of the above problems


Submission

A paper submission has to be in English, in pdf format, and at most 8 pages (excluding references) in CVPR style. The paper format must follow the same guidelines as for all CVPR submissions.
https://cvpr2022.thecvf.com/author-guidelines
The review process is double blind. Authors do not know the names of the chair/reviewers of their papers. Reviewers do not know the names of the authors.
Dual submission is allowed with CVPR main conference only. If a paper is submitted also to CVPR and accepted, the paper cannot be published both at the CVPR and the workshop.

For the paper submissions, please go to the online submission site
https://cmt3.research.microsoft.com/MAI2022

Accepted and presented papers will be published after the conference in the CVPR Workshops Proceedings on by IEEE (http://www.ieee.org) and Computer Vision Foundation (www.cv-foundation.org).

The author kit provides a LaTeX2e template for paper submissions. Please refer to the example for detailed formatting instructions. If you use a different document processing system then see the CVPR author instruction page.

Author Kit: https://cvpr2022.thecvf.com/sites/default/files/2021-10/cvpr2022-author_kit-v1_1-1.zip

Workshop Dates

● Regular Papers Submission Deadline: March 20, 2022
● Decisions: April 14, 2022
● Camera Ready Deadline: April 19, 2022



MAI 2022 (jointly with AIM workshop @ ECCV2022) has the following associated groups of challenges (ongoing!):

● Video Super-Resolution (Evaluation platform: MediaTek Dimensity APU) - Powered by MediaTek
● Image Super-Resolution (Eval. platform: Synaptics Dolphin NPU) - Powered by Synaptics
● Learned Smartphone ISP (Eval. platform: Snapdragon Adreno GPU) - Powered by OPPO
● Bokeh Effect Rendering (Eval. platform: ARM Mali GPU) - Powered by Huawei
● Depth Estimation (Eval. platform: Raspberry Pi 4) - Powered by Raspberry Pi


PARTICIPATION

To learn more about the challenges and to participate:
https://ai-benchmark.com/workshops/mai/2022/


Challenges Dates

● Release of train data: 24th of May, 2022
● Validation server online: end of May, 2022
● Competitions end: end of July, 2022
● Challenge paper submission deadline (AIM workshop @ ECCV2022): 10th of August, 2022


Organizers (TBU)


● Andrey Ignatov ( ETH Zurich)
● Radu Timofte ( University of Wurzburg and ETH Zurich)
● Luc Van Gool ( KU Leuven and ETH Zurich)
● Cheng-Ming Chiang ( MediaTek Inc.)
● Hsien-Kai Kuo ( MediaTek)
● Kim Byeoung-su ( Samsung Electronics Co., Ltd.)
● Gaurav Arora ( Synaptics Inc.)
● Abdel Younes ( Synaptics Inc.)
● David Plowman ( Raspberry Pi (Trading) Ltd.)
● Eirikur Agustsson ( Google)
● Chiu Man Ho ( OPPO)
● Zibo Meng ( OPPO)
● Shuhang Gu ( University of Sydney and OPPO)


Speakers (TBA)


Website: https://ai-benchmark.com/workshops/mai/2022/