marc davis chairman and chief technology officer representing video for retrieval and repurposing...

44
Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

Post on 22-Dec-2015

214 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

Marc DavisChairman and Chief Technology Officer

Representing Videofor Retrieval and Repurposing

SIMS 202 Information Organization and Retrieval

Page 2: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Presentation Outline

• Introductions

• Problem space and motivation

• Current approaches

• Issues in video representation and retrieval

• Media streams demonstration

Page 3: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Global Media Network

• Digital video produced anywhere by anyone accessible to anyone anywhere

• Today’s video users become tomorrow’s video producers

• Not 500 Channels — 500,000,000 Video Web Sites

Page 4: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

What is the Problem?

• Today people cannot easily create, find, edit, share, and reuse media

• Computers don’t understand video content– Video is opaque and data rich– We lack structured representations

• Without content representation (metadata), manipulating digital video will remain like word-processing with bitmaps

Page 5: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

The Search for Solutions

• Current approaches don’t work– Signal-based analysis– Keywords– Natural language

• Need standardized metadata framework– Designed for video and rich media data– Human and machine readable and writable– Standardized and scaleable– Integrated into media capture, production, editing,

distribution, and reuse– Enables widespread use and reuse of video in daily life

Page 6: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Signal-Based Parsing

• Theoretical problem– Mismatch between percepts and concepts

(e.g., dogs, friends, cars)

• Practical problem– Parsing unstructured, unknown video is very,

very hard

Page 7: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Signal-Based Parsing

• Some things are doable and usable– Video

• Scene break detection• Camera motion• Low level visual similarity

– Audio• Pause detection• Audio pattern matching• Simple speech recognition

• Some things can be made easier – At the point of capture, simplify and/or interact with the recording

device, the environment, and agents in the environment– If not, after capture use “human-in-the-loop” algorithms

Page 8: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Keywords vs. Semantic Descriptors

dog,biting,Steve

Page 9: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Keywords vs. Semantic Descriptors

dog,biting,Steve

Page 10: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Why Keywords Don’t Work

• Are not a semantic representation

• Do not describe relations between descriptors

• Do not describe temporal structure

• Do not converge

• Do not scale

Page 11: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Natural Language vs. Visual Language

Jack, an adult male police officer, while walking to the left, starts waving with his left arm, and then has a puzzled look on his face as he turns his head to the right; he then drops his facial expression and stops turning his head, immediately looks up, and then stops looking up after he stops waving but before he stops walking.

Page 12: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Natural Language vs. Visual Language

Jack, an adult male police officer, while walking to the left, starts waving with his left arm, and then has a puzzled look on his face as he turns his head to the right; he then drops his facial expression and stops turning his head, immediately looks up, and then stops looking up after he stops waving but before he stops walking.

Page 13: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Visual Language Advantages

• A language designed as an accurate and readable representation of video (especially for actions, expressions, and spatial relations)

• Enables Gestalt view and quick recognition of descriptors due to designed visual similarities

• Supports global use of annotations

Page 14: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Representing Video

• Streams vs. Clips

• Video syntax and semantics

• Ontological issues in video representation

• Retrieving video

Page 15: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Video is Temporal: Streams vs. Clips

Stream of 100 Frames of Video

A Clip from Frame 47 to Frame 68 with Descriptors

Page 16: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Streams vs. Clips

• Clip-based representation

–Fixes a segmentation of the video stream

–Separates the clip from its context of origin

–Encodes only one particular segmentation of

the original data

Page 17: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Streams vs. Clips

The Stream of 100 Frames of Video with 6 Annotations Resulting in ManyPossible Segmentations of the Stream

Stream of 100 Frames of Video

Page 18: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Streams vs. Clips

• Stream-based representation

–The stream of frames is left intact

–The stream has many possible segmentations by multi-layered annotations with precise time indexes (and the intersections, unions, etc. of these annotations)

Page 19: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Stream-Based Representation

• Makes annotation pay off– The richer the annotation, the more numerous the

possible segmentations of the video stream

• Clips – Change from being fixed segmentations of the video

stream, to being the results of retrieval queries based on annotations of the video stream

• Annotations– Create representations which make clips, not

representations of clips

Page 20: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Video Syntax and Semantics

• The Kuleshov Effect

• Video has a dual semantics

– Sequence-independent invariant semantics of shots

– Sequence-dependent variable semantics of shots

Page 21: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Ontological Issues for Video

• Video plays with rules for identity and continuity

– Space

– Time

– Character

– Action

Page 22: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Space and Time

• Actual Recorded Space and Time– GPS– Studio space and time

• Inferable Space and Time– Establishing shots– Cues and clues

Page 23: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Time: Temporal Durations

• Story (Fabula) Duration– Example: Brushing teeth in story world (5 minutes)

• Plot (Syuzhet) Duration– Example: Brushing teeth in plot world (1 minute: 6

steps of 10 seconds each)

• Screen Duration– Example: Brushing teeth (10 seconds: 2 shots of 5

seconds each)

Page 24: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Character and Continuity

• Identity of character is constructed through– Continuity of actor– Continuity of role

• Alternative continuities– Continuity of actor only– Continuity of role only

Page 25: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Representing Action

• Describe the intersubjective, physically visible aspects of what you see and hear– Emotions vs. expressions– Abstract actions vs. conventionalized actions

• Consider how actions can be decomposed and combined (temporally and spatially) – Actions and subactions

• Consider how actions can be recontextualized– By montage and reuse– By cultural differences

Page 26: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Retrieving Video

• Query:

– Retrieve a video segment of “a hammer hitting a nail into a piece of wood”

• Sample results:

– Video of a hammer hitting a nail into a piece of wood

– Video of a hammer, a nail, and a piece of wood

– Video of a nail hitting a hammer, and a piece of wood

– Video of a sledgehammer hitting a spike into a railroad tie

– Video of a rock hitting a nail into a piece of wood

– Video of a hammer swinging

– Video of a nail in a piece of wood

Page 27: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Types of Video Similarity

• Semantic– Similarity of descriptors

• Relational– Similarity of relations among descriptors in

compound descriptors

• Temporal– Similarity of temporal relations among descriptors

and compound descriptors

Page 28: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Retrieval Examples to Think With

• “Video of a hammer, a nail, and a piece of wood”– Exact semantic and temporal similarity, but no relational similarity

• “Video of a nail hitting a hammer, and a piece of wood” – Exact semantic and temporal similarity, but incorrect relational similarity

• “Video of a sledgehammer hitting a spike into a railroad tie”– Approximate semantic similarity of the subject and objects of the action and

exact semantic similarity of the action; and exact temporal and relational similarity

• “Video of a hammer swinging” cut to “Video of a nail in a piece of wood”– Combines two disparate elements in the database (partial results) to create

an effective query response

Page 29: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Media Streams

Page 30: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Media Annotation and Retrieval Engine

• Key benefits– More accurate annotation and retrieval– Global usability and standardization– Reuse of rich media according to content and structure

• Key features– Stream-based representation (better segmentation)– Semantic indexing (what things are similar to)– Relational indexing (who is doing what to whom)– Temporal indexing (when things happen)– Iconic interface (designed visual language)– Universal annotation (standardized markup schema)

Page 31: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Media Streams Demonstration

Page 32: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Media Streams GUI Components

• Media Time Line

• Icon Space– Icon Workshop– Icon Palette

Page 33: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Media Time Line

• Visualize video at multiple time scales

• Write and read multi-layered iconic annotations

• One interface for annotation, query, and composition

Page 34: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Media Time Line

Page 35: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Icon Space

• Icon Workshop– Utilize categories of video representation– Create iconic descriptors by compounding iconic

primitives– Extend set of iconic descriptors

• Icon Palette– Dynamically group related sets of iconic descriptors– Reuse descriptive effort of others– View and use query results

Page 36: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Icon Space

Page 37: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Icon Space: Icon Workshop

• General to specific (horizontal)– Cascading hierarchy of icons with increasing

specificity on subordinate levels

• Combinatorial (vertical)– Compounding of hierarchically organized

icons across multiple axes of description

Page 38: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Icon Space: Icon Workshop Detail

Page 39: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Icon Space: Icon Palette

• Dynamically group related sets of iconic descriptors

• Collect icon sentences

• Reuse descriptive effort of others

Page 40: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Icon Space: Icon Palette Detail

Page 41: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Video Retrieval In Media Streams

• Same interface for annotation and retrieval

• Assembles responses to queries as well as finds them

• Query responses use semantics to degrade gracefully

Page 42: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Media Streams Technologies

• Minimal consensual representation distinguishing video syntax and semantics

• Iconic visual language for annotating and retrieving video content

• Retrieval-by-composition methods for repurposing video

Page 43: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Methodological Considerations

• Techne-centered methodology combines– Constructing theories by constructing artifacts– Constructing artifacts informed by (de)constructing

theories– Practitioners

• Lev Kuleshov, Sergei Eisenstein, Seymour Papert, Narrative Intelligence Reading Group, Marc Davis

• Designing video representation and retrieval systems requires a techne-centered methodology

Page 44: Marc Davis Chairman and Chief Technology Officer Representing Video for Retrieval and Repurposing SIMS 202 Information Organization and Retrieval

amova

Amova Proprietary

Future Work

• MPEG-7 standardization efforts• Gathering more and better metadata at the point

of capture• Integrating metadata into conventional media

editing and sharing• Developing “human-in-the-loop” indexing

algorithms and systems• Representing action sequences and even

higher level narrative structures• Fair use advocacy