Browse Prior Art Database

Tracking Mickey Mouse and Other Cartoon Characters

IP.com Disclosure Number: IPCOM000123778D
Original Publication Date: 1999-Apr-01
Included in the Prior Art Database: 2005-Apr-05
Document File: 1 page(s) / 58K

Publishing Venue

IBM

Related People

Coden, AR: AUTHOR

Abstract

In this paper we disclose a novel way of automatically detecting cartoon characters in animated movies (videos). Furthermore, characteristics like movement, gait and gestures can be automatically found to. These features are important, as they are repeated many times within an animated feature as they give a cartoon character its "character". Some of the techniques could lead also to a compression algorithm in terms of how many pixels need to be stored to effectively be able to reuse a particular sequence of actions of a character.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 55% of the total text.

Tracking Mickey Mouse and Other Cartoon Characters

   In this paper we disclose a novel way of automatically
detecting cartoon characters in animated movies (videos).
Furthermore, characteristics like movement, gait and gestures can be
automatically found to.  These features are important, as they are
repeated many times within an animated feature as they give a cartoon
character its "character".  Some of the techniques could lead also to
a compression algorithm in terms of how many pixels need to be stored
to effectively be able to reuse a particular sequence of actions of a
character.

   Cartoon Characters like Mickey Mouse or Bugs Bunny for
example have very distinctive features and outline shapes.  Mickey
Mouse's round ears for instance can be spotted quite easily as their
outline is always round independent of the perspective he is drawn.
The outline of Bug's Bunny's ears are simple transformation of the
function describing the outline of a frontal drawing of him.  Not
only is the outline of a comic strip character pretty easily
detectable for a human, but so are other characteristics like voice,
gait and gestures.  We suggest to build an automatic system which can
detect these characteristics and log them.  The purpose is manyfold.

   Movies and videos are increasingly produced or
transformed into digital format and stored in a video database.
Powerful search engines are part of such database to be able to
retrieve particular sequences from a video.  Howev...