Video Scene Detection

Daniel Rotman

Abstract

Video scene detection is the task of dividing a video into semantic sections. I will present our novel and effective method for temporal grouping of scenes using an arbitrary set of features computed from the video. This task is formulated as a general optimization problem and an efficient solution is provided using dynamic programming. Our unique formulation allows us to directly obtain a temporally consistent segmentation, unlike many existing methods, and has the advantage of being parameter-free. I will also present how we expanded the method to incorporate features from multiple modalities, and I will present a novel technique to estimate the number of scenes in the video using Singular Value Decomposition (SVD) as a low-rank approximation of a distance matrix. This method proved to perform outstandingly and resulted in three published papers.

Speaker

Photo of Daniel Rotman

Daniel Rotman is a research scientist in the Video and GIS Analytics group at IBM Research - Haifa. Daniel received his B.Sc and M.Sc. at the Technion in the Electrical Engineering department.

Lecture languages

EnglishHebrew

Topics

AI / Automation

Duration options

1 hour

Travel/delivery options

In-countryOutside of country: Open for discussionRemote via video conference

Country

Israel

Lecture booking request

Thank you for your interest in hosting an IBM speaker. Please fill out the following form with as much detail as possible. An IBM representative will reach out to discuss your booking request. All guest lectures are subject to availability and agreements under this collaboration are not legally binding.