Browse Prior Art Database

Temporally-ordered multi-dimensional attribute visualization and similarity measure of video scenes

IP.com Disclosure Number: IPCOM000244579D
Publication Date: 2015-Dec-23
Document File: 2 page(s) / 43K

Publishing Venue

The IP.com Prior Art Database

Abstract

We propose a system to solve the problem of how to visualize and compare video clips, with focus on the temporal flow of video clips with multi-dimensional attributes. An example for such multi-dimensional attributes of a video can be a video of a scientific experiment where the multi-dimensional attributes are 'Classroom', 'Lab', 'Outdoor', 'Computers' and 'People'.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

Temporally- video scenes

We propose a system to solve the problem of how to visualize and compare video clips , with focus on the temporal flow of video clips with multi-dimensional attributes. An example for such multi-dimensional attributes of a video can be a video of a scientific experiment where the multi-dimensional attributes are 'Classroom', 'Lab', 'Outdoor', 'Computers' and 'People'.

More broadly, as video repositories become larger, it becomes more and more difficult to inspect their contents efficiently. Moreover, videos are typically non-homogeneous and are composed of different, heterogeneous scenes. Inspecting the contents of the different scenes in a video is therefore also a time-consuming task. Also, the temporal ordering of the different scenes in a video is also an important factor for comparison between videos.

A related example scenario is video search, in which a result set of videos is presented (as a response to the search), and the need is to represent the videos in a user friendly way so it's easy to capture quickly -- using an effective visualization -- the essence of each video in the result set. Pattern finding is also a related task, in which a user defines a combination of scenes by a set of attributes and temporal order and the system returns matching videos. For example, a user may look for a video that begins in an indoor environment and then transitions to an outdoor environment, and the system will return a list of videos that follow this pattern.

    The proposed system simultaneously encodes (by visualization) the temporal sequence of video scenes as a flow, the multi-dimensional attributes as a pattern, and in addition the concentration of patterns with respect to the attributes. In addition to the afore...