Low-Power Computer Vision Challenge

2020 CVPR Workshop

Call for Paper

Seattle, USA

Date: 2020/06/15 8:30AM-5:00PM

Call for Papers (Three Tracks)

  1. Solutions for Low-Power Computer Vision
  2. Grand Challenges in Low-Power Computer Vision
  3. Poster for Low-Power Computer Vision

Track 1: Solutions for Low-Power Computer Vision:

Computer vision technologies have made impressive progress in recent years, but often at the expense of increasingly complex models needing more and more computational and storage resources. This workshop aims to improve the energy efficiency of computer vision solutions for running on systems with stringent resource constraints, for example, mobile phones, edge devices, Internet of Things, drones, or renewable energy systems. Efficient computer vision can enable many new applications (e.g., wildlife observation) powered by ambient renewable energy (e.g., solar, vibration, and wind). This workshop will discuss the state of the art of low-power computer vision, challenges in creating efficient vision solutions, promising technologies that can achieve the goals, methods to acquire and label data, benchmarks and metrics to evaluate progress and success. Authors are encouraged to present innovation in any part of the entire system, such as new hardware components, new algorithms, new methods for system integration, new semiconductor devices, and new computing paradigms. This workshop emphasizes “system-level” solutions with implementations for demonstrations and experiments. Conceptual designs or solutions for individual components without integration into functional systems are discouraged.

Authors are encouraged to discuss the following issues in their papers:

  • Description of the vision problem
  • Description of the solution
  • Description of the connections between low-power and computer vision
  • Description of the data for evaluation
  • Description of the system requirements (for example, but not limited to, solar powered, autonomous robots, drones)
  • Description of the entire system
  • Metrics for evaluation
  • Comparison with the state-of-the-art

Track 2: Grand Challenges in Low-Power Computer Vision:

Competitions are widely adopted in academia, industry, and government to propel technology forward. Well-known competitions include the DARPA Autonomous Vehicle Challenge and the Space X prize. The competitions have been attributed as the accelerators of making significant progress in the technologies. This workshop solicits papers that describe future competition in low-power computer vision. Authors are encouraged to think boldly, imagining competitions a la “Grand Challenge” or “X Prize”. The winners’ solutions (hardware, software, or some combination) should be far beyond today’s available technologies.

Authors are encouraged to discuss the following issues in their papers:

  • Description of the challenge and why it is worth the community’s efforts for several years
  • Description of the state-of-the-art (the challenge should be far beyond today’s technology but achievable)
  • Description of the connections between low-power and computer vision
  • Description of how people that can benefit from the technologies
  • Description of the market size in billions of dollars per year (if applicable)
  • Description of the data (e.g., image, video, with or without captions, etc.)
  • Methods to acquire the data and methods for scoring
  • Paths toward solving the problems. Do they need innovation in vision algorithms? Hardware? Software? Will non-traditional systems be needed? Why?
  • Description about copyright and privacy (if applicable)
  • Methods to annotate the data (if applicable, for getting the ground truth)
  • Constraints of hardware (if applicable, e.g., weight and size)
  • Minimum or maximum requirements (if applicable, e.g., accuracy and execution time)

Track 3: Posters for Low-Power Computer Vision:

The poster track is designed for projects that can benefit from an interactive setting between authors and audience. Authors may choose this format to present projects that may have, for example, demonstrations or user participation. Students may take advantage of this format and solicit suggestions or critiques of their projects. Researchers and practitioners from industry may present their projects and recruit users. The topics can be on either solutions for low-power computer vision or grand challenges in low-power computer vision. Works-in-progress are welcome.

Authors are encouraged to discuss the following issues in their papers:

  • Description of the vision problem
  • Description of the low-power solution
  • Description of the connections between low-power and computer vision
  • Description of the stage of this project
  • Description of the demonstration (if applicable)

Submission

All submissions must be in PDF format. Submissions are limited to eight pages (tracks 1 and 2) or two pages (track 3), including all figures and tables. References are not included in the page limit. All submitted papers will be double-blind reviewed. Authors must not reveal their identities. Please refer to the Author Guidelines (http://cvpr2020.thecvf.com/submission/main-conference/author-guidelines#call-for-papers)

The papers selected from Tracks 1 and 2 may be in the form of oral presentations or posters (in the event of a large number of outstanding papers). The papers selected from Track 3 will be in the form of posters. If a paper in Track 1 or 2 is selected for a poster, the authors will need to shorten the paper to two pages.

Important dates (Anywhere On Earth)

  • Online paper submission: 2020/03/27 (will not be extended)
  • Notification to authors: 2020/04/10
  • Final version: 2020/04/17 (This is a deadline set by CVPR)

Submission site: https://cmt3.research.microsoft.com/LPCV2020

Student Travel Grants

Travel grants will be available to students. Details will be available in April.

Sponsors

Organizations interested in sponsoring the workshop, please contact Dr. Yung-Hsiang Lu, yunglu@purdue.edu. Check the all sponsors at homepage Sponsors