Big Data Summarization for Computer Vision

Dear Colleagues,

The increase in both quantity and quality of sensors, and specifically cameras, in the past decade has made huge streams of high-resolution visual data available. Such data can be utilized for localization, mapping, scene understanding, 3D modeling, and training neural networks, just to name a few. However, to leverage big data, one must overcome a series of challenging tasks, including efficient (and possibly parallel) computation, efficient storage, and tackle streaming and dynamic data.

While the common approach is to design faster and more efficient algorithms, an alternative and modern approach is to provably approximate the big data via a much smaller data. Such compression schemes are often called a coreset or a sketch. Running existing (possibly inefficient) algorithms on reduced (small) data requires less time, space, and energy and would produce a result that is provably close to the solution obtained from running on complete (big) data.

The aim of this Special Issue, “Big Data Summarization for Computer Vision”, is to attract original and innovative research results which successfully manage to reduce the computational and storage overload of computer vision tasks via designated compression and approximation schemes.

Topics of interest include but are not limited to:

• Compressions and approximations for pose estimation;
• Fast and accurate simultaneous localization and mapping (SLAM) via coresets and sketches;
• Coresets and sketches for 3D reconstruction and modeling;
• Faster training of deep learning models via dataset compression;
• Compressing deep learning models via coresets and sketches.

In addition to application-driven contributions, this Special Issue also welcomes submissions with a deep theoretical and provable aspect. These papers will both act as a guide and encourage joint works between theoretical researchers and practitioners, aiming to develop provable algorithms for real-world computer vision problems.

Dr. Ibrahim Jubran
Prof. Dr. Dan Feldman
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Information is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 1600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• Computer vision
• Localization and mapping
• Compression of big data
• Coresets
• Sketches