Photo Metadata Standards

Author: Raphaël Troncy

$Date: 2007/09/13 12:15:38 $, $Revision: 1.9 $


Leesklub on Fri 06/07/2007

The revision 1.6 of this blue note has been discussed during the leesklub. The discussion focuses on what are our research questions given that there is a strong will for having more "semantics" metadata for photos.
   - What do we want to do with these semantics metadata?
   - How can we get and infer rich semantics metadata?

Interfaces could/should group the resulting images of a query, displaying a representative image of each cluster. The grouping could use the semantics of the annotation properties + some knowledge of the world.

Use Cases:

TO DO:


Table of Contents


1. Introduction

The starting point of this blue note is the IPTC Photo Metadata White Paper 2007 [1] and the Photo Metadata Conference held the 7th of June 2007 in Florence (Italy). This blue note aims at summarizing the current initiatives but also describing the open issues for representing and exchanging metadata of photographs.

Metadata are fundamental:

2. Photo Metadata Conference

Cepic hosted the first international Photo Metadata Conference entitled "Working Towards A Seamless Photo Workflow". The conference was jointly organised by the standards organisation IPTC and media organisation IFRA, and sponsored by Adobe with the aim of bringing together the many and varied parties that have an interest in improving the application of metadata standards to digital images. The conference attracted a heavyweight audience of 130 delegates.
Implicit goal: avoid the "culture of blame" that can prevail – and work together to resolve the issues.

2.1 Keynote - Andreas Trampe, head of the photodesk of "Stern"

General problem: Too many photographs and very badly indexed!

2.2 Report from Metadata Users

Among the image creators were two photographers well known for representing the interests of freelance photographers in search of better standards, David Riecks and Peter Krogh. Among the picture users were the Head of the Stern Photodesk, Andrea Trampe, Roger Bacon of Reuters, Simon Span of Trinity Mirror.

2.3 Report from Photo Metadata Standardisation Bodies

Representatives of the professional and standards bodies included Harald Toeffler of IFRA, Michael Steidl of IPTC, Jeff Sedlik of PLUS.

2.4 Report from Photo Metadata Implementers

Manufacturers and software developers for presentations of some potential solutions to the problems aired earlier in the day – namely Gunar Penikis of Adobe Systems Inc; Joe Schorr of Apple Inc.; Clemens Molinari of Fotoware; and Josh Weisberg of Microsoft, with Peter Stig of Hasselblad and Hiroshi Maeno of Canon representing the camera manufacturers.

"The Truth Is In The File" Josh Weisberg (Microsoft)!
Microsoft has created what they call WIC (Windows Imaging Components). WIC is an extensible, system-wide platform for image handling that will reduce the need for sidecar files. It ships with Vista and is available for Windows XP and other server products. This makes it possible for developers to provide support for multiple metadata standards in their application. Out of the box, the WIC policy handler can support EXIF, IPTC and XMP.

3. Photo Metadata White Paper

This white paper [1] is co-authored by numerous persons from the IPTC Photo Metadata Working Group and representing: news agencies, PLUS, CEPIC, IPTC, Newspapers, Stock Photo Agency, Alinari (participant in the SemaPhoto proposal), etc.
A mailling list has been created on an IPTC moderated Yahoo forum: http://groups.yahoo.com/group/iptc-photometadata in which I'm subscribed.
Comments and contributions on this white paper are expected to be sent to office@iptc.org.

Goals or rather slogan: PRESERVE METADATA Metadata properties can be grouped into four categories:

Descriptive Metadata:

Property nameProperty descriptionValue
headline A brief summary of the contents of the photograph.
Note: The headline term should not be confused with title.
Free text
description/caption Prose to comprehensively describe the who, what, when, why and where of the photograph.
Note: People in the photo must be placed into the person field. Organisations or companies shown in some way on the photo must be placed into the organisation field.
Free text
location (in picture) Defines the location shown in the photograph.
Note: Location may be defined by hierarchical terms representing the world region, country, province/state, city or any location outside a city.
Controlled vocabulary: TGN?
person (in picture) Name of the known person(s) in the photograph.
Note: Value should be split into surname and given name.
Controlled vocabulary
additional model information Information about the age and ethnicity of the model(s) in the photograph. Free text
organisation (in picture) Name of the organisation or company in the photograph. May be supplemented by values, such as stock ticker symbols, from a controlled vocabulary. Controlled vocabulary
emotion Describes the emotion shown in the photo, if people are pictured – should be a value from a controlled vocabulary. Controlled vocabulary
keywords Keywords to express the subject of the content in free text.
Note: Values from controlled vocabularies should be placed in such corresponding properties as Subject NewsCodes or Scene NewsCodes or person, organisation, or location.
Free text
Subject NewsCodes Describes the most prominent subjects of the photo by one or more codes from the IPTC "Subject NewsCodes". IPTC News Codes
Scene NewsCodes Describes the scene of a photo content by one or more codes from the IPTC "Scene NewsCodes". IPTC News Codes
genre Describes the nature of the photo. This may be journalistic, artistic or intellectual characteristics. ???
predominant colour A value for the visually predominant colour of the photo. The value should be from a controlled vocabulary of about 20 typical colours. Controlled vocabulary

Administrative Metadata:

Property nameProperty descriptionValue
unique id The globally unique identifier (GUID) valid in the context of the delivering provider and at the time of the delivery. URI
other id(s) An identifier, preferably a GUID, of a photo or another news object which relates to this photo. This property should be complemented by a qualifier expressing the kind of relationship; it may take values like e.g. “previous”, “see also”, or “similar”. URI
title A shorthand reference for the image or photo – primarily for an identification less formal than the unique id.
Note: The title of an image may take several forms; for photographers, this might be the filename of their original scan or digital camera file.
Free text
date created Records the date (and optionally the time) when the photograph was taken, not the date when scanned, imported, archived or edited. This field should support a datatype for dates that allows one to express a date range, so this may be used for historic photographs or other pictures with creation dates that are not known exactly. Date
location created Defines the location where the photograph was taken.
Note: This may be different from the location of the subject of the photo, such as in long distance shots.
Controlled Vocabulary
event A free text field naming or describing the specific event at which the photo was taken, e.g. a press conference, dedication ceremony, etc. Free text
job id Identifier for the purpose of improved workflow handling. This ID should be added by the creator or provider for transmission and routing purposes only and should have no significance for archiving. ???
instructions Any of a number of instructions from the provider or creator to the receiver of the photo, not covered by other explicit fields. May cover e.g. embargoes, instruction for reproduction, etc. Free text
description/caption writer Name of the person(s) involved in writing, editing or correcting the caption or other metadata of the photo. Free text

Rights Metadata:

Property nameProperty descriptionValue
creator This should preferably be the person who created the photo. If this is not appropriate then the name of an organisation or company. Free text
creator job title The job title of the photographer, such as "Staff Photographer", "Freelance Photographer" or "Independent Commercial Photographer". Enumerated list
credit line One or more parties to be credited for this photograph; this single line is intended to be shown adjacent to a photo. Essentially free text, this line may follow formatting and style guides as required by law or company policy. Free text
copyright notice A free text notice containing any necessary copyright notice for claiming the intellectual property, which should identify the current owner of the copyright for the photograph. Free text
license contact The person or company that should be contacted for licensing the image, if there are any open issues regarding copyright or usage. Free text
model release A value indicating whether “none”, “unlimited” or “partial” model releases are available for this photograph, or whether they are applicable. An optional set of "release identifiers" should refer to corresponding release documents being accessible over the internet. Enumerated list
property release A value indicating whether “none”, “unlimited” or “partial” property releases are available for this photograph, or whether they are applicable. An optional set of "release identifiers" should refer to corresponding release documents being accessible over the internet. Enumerated list
other third party rights A free-text field with statements regarding all other third party rights not covered by the Model Release and Property Release fields, such as paintings or images featured prominently in the photo. Free text
usage rights The usage rights field should include text instructions on how the photograph can be legally used. It is strongly encouraged that you use a standardized set of terms or controlled vocabulary when populating this field, or consider a reference to a PLUS code. Free text
provider Who provided the photograph to the receiver. This doesn't have to be the owner/creator. Free text

Technical Properties Metadata:

Property nameProperty descriptionValue
colour space The current colour space of the photo such as sRGB, Adobe RGB, etc. Enumerated list
orientation The current orientation of the photo such as “landscape” or “portrait” Enumerated list
max available height+width Maximum available size of the image in pixels; the high resolution rendition may be a different image. Number
physical type of original One of the following terms “film”, “slide”, “digital image”, etc. Enumerated list

4. Open issues

  1. Identifiers:
  2. Interoperability:
  3. Metadata trust:
  4. Technical metadata:

One of the most important new initiatives to emerge from the conference was Canon’s announcement that it is working to develop a permanent image ID that is created at the time the image is created and cannot be removed.

The new release of the Adobe’s XMP toolkit is open-source and therefore freely available for adaptation by specialist users.
XMP can hold multiple forms of metadata (EXIF, IIM, IPTC Core) within a single container, and it is possible to extend it – by allowing other applications to edit XMP without fully understanding the file format. XMP can support formats beyond still images such as AVI, WAV, MOV, MP3, and MPEG.
It is to build and install custom metadata panels for Adobe Photoshop and Bridge. Bridge can now also be extended with Flash panels, which can have a nice user interface and via scripting be extended to send images through FTP or web services.
Additional details can be found at the following URL’s:

5. Use Cases

5.1 Tony Blair

Scenario:
You have been asked to do a portrait of Tony Blair. As this is the last time he’ll be attending a G8 summit meeting, it is a good time to review his ten years in power. This search is a little more difficult to limit in terms of the time frame, because you need images from the past ten years of his premiership. What were the highlights of his ten years in power?
It’s early in the morning, you haven’t had your cup of coffee yet and the only notable events you can think of are that he was an adamant supporter of the Northern Ireland peace process, that he sent his soldiers to fight in Iraq and that the British economy is doing quite well.

Alia's suggestions:
  1. See an overview of his carrear: preferably in historical order.
  2. Narrow the results into different perspective when there are too many image results:
  3. I need a background on Tony Blair as a person, not as merely as a political icon.

6. References

  1. IPTC. IPTC Photo Metadata White Paper 2007, version 11, 2007.
  2. Raphaël Troncy, Jacco van Ossenbruggen, Jeff Z. Pan and Giorgos Stamou. Image Annotation on the Semantic Web, W3C Multimedia Semantics XG Report, 2007.
  3. Susanne Boll. Photo Use Case, W3C Multimedia Semantics XG, 2007.