This workshop focuses on resource efficient deep learning for computer vision, and aims to bring together researchers and practitioners from academia and industry to share their expertise, ideas, and best practices on this important topic. The goal of this workshop is to provide a platform for the exchange of ideas and discussion on the latest advancements in resource efficient deep learning for computer vision, and to identify future directions for research and development in this area. 

Topics of interest for this workshop include, but are not limited to:

• Designing efficient neural architectures

• Compression of existing models

• Efficient fine-tuning of large-scale vision and language models.

• Deep learning on very large images

• Energy-efficient deep learning hardware and accelerators

• Benchmarking datasets for model efficiency

• Efficient processing of videos in computer vision

• Federated learning, distributed learning

TRAVEL GRANT: We will provide a travel grant to support a student author from a developing country. This travel grant will only be applicable for the long paper category and will be capped to a maximum of 5500 NOK (~520 USD). The travel grant is sponsored by Norwegian Artificial Intelligence Research Consortium (NORA).

Submission Instructions

Submission link: https://cmt3.research.microsoft.com/RCV2023

All papers need to be formatted as per the ICCV 2023 guideliness. Please use the template made available at iccv2023AuthorGuidelines. 

Papers can be submitted in two different formats: 

• Long paper: Long papers should not be not more than 8 pages in length excluding references. All the requirements are the same as the ICCV 2023 paper style guide. Long papers are meant to discuss mature and thoroughly investigated work. A long paper should not only describe novel ideas but also have full experiments and analyses to support the proposed ideas.

• Short paper: Short papers should not exceed 4 pages excluding references. All other requirements are the same as the ICCV 2023 paper style guide. Short papers are meant to discuss ideas that are at an early stage and only partly investigated. The presented idea should be novel and the paper is expected to contain the basic experiments supporting the claim.l Comprehensive experiments and. analyses, however, are not necessary.

Note that the submission for both types of papers is using the same portal. Papers less than 4 pages will be reviewed as short papers, those more than 4 pages  (excluding references) will be reviewed as long papers, and papers beyond 8 pages in length (excluding references) will be desk rejected.

All the papers to be submitted through the Conference Management Toolkit (CMT) link: https://cmt3.research.microsoft.com/RCV2023