Copyright Lincoln Cushing, Docs Populi http://www.docspopuli.org
Originally written as student paper at UC Berkeley SIMS 11/21/2000, updated 3/5/2007
This is a subject in perpetual change; please let author know of any changes, corrections, or suggestions.
Protection of Digital Images on the Web
“There is agreement among academics, educators, and librarians alike about the rationale for creating ... a national digital library to provide wider electronic access to knowledge and information. Such access would (1) vastly enhance education, scholarship, arts, and science; (2) preserve and make accessible unique primary materials about our heritage; (3) make America more competitive in world markets; (4) foster an informed citizenry.”
“The National Digital Library Program—A Library for All Americans,” The Library of Congress, February 1995, p. 5. 
“Janus from Brussels has told me that he has e-mailed your posters to a group of college students who in turn have altered them with the help of an Adobe graphics program and they have successfully turned their messages around… The students are also creating a counter web site to showcase the posters and to enlighten the international public about the horrors of communism. Janus told me that the students have a message for you of course translated from the Flemish which reads “What are you going to do to us? Sue us? WE CAN SAY WE TOOK THE POSTERS FROM THE PICTURE LIBRARY AT SCHOOL. YOU HAVE NO PROOF.”
e-mail from Serge Goldberg to the Cuba Poster Project message board July 10, 2000; emphasis exactly as written. 
Any sharing or display of digital images on the World Wide Web poses a risk for inappropriate copying and use. Given the ease with which material can be swiped off the screen and the huge numbers of people clicking through websites, is it reasonable and appropriate for institutions such as libraries, galleries, and museums to post anything at all? I would suggest that the answer is a qualified yes; the benefits far outweigh the dangers if the hosting institution truly understands the issues and diligently adopts a methodology for protecting digitally-published images. The issues involved transcend simple protection of the artifact, and ultimately include the broad challenge of protecting the integrity of the intellectual content.
Arguments for digitizing and publishing images on the Web
There are many good reasons to put images up on the Web. From the viewpoint of simple access, some oversized documents, such as posters, are difficult for institutions to store in a way that allows access for researchers.  Many materials are fragile and suffer when handled, even by professional staff during routine operations. Institutions may be inaccessible to the physically disabled or geographically remote for those that would like to see their holdings. Perhaps the most profound development is the tremendous advantage that properly metadata-tagged digital images offer for computer-assisted research. It is finally beginning to be possible to rapidly assemble whole bodies of work where the original artifacts reside in a multiplicity of physical locations. Institutions have chosen to approach document digitization to meet a variety of needs, including document conservation and document surrogates. The practical contradiction emerges as being the tradeoff between high resolution (at high cost and more taxing on storage devices, but with more options for use) and moderate resolution (at lower cost but with more limited uses). It is outside the scope of this paper to address these tradeoffs, but at some point all image publishers will confront the same issues regarding protection of digital images.
Why protect works at all?
The legal and ethical issues surrounding the notion of intellectual property itself are beyond the scope of this paper. However, it must be noted that much of this is deeply contested terrain. At one end of the spectrum are those that believe strongly that without rigorous laws protecting human expression as property that intellectual creativity will wither through lack of market incentives. At the other end are those that consider all property theft, that acts of personal creativity are based on the uncredited works of others, and that the commodification of creative works has distorted the entire cultural landscape. This debate has given rise to various approaches to crediting and describing materials, ranging from conventional patents and copyrights to “copyleft”  and deliberate description of materials as being copyright-free.
Related to these issues is the concept of “fair use”, whereby under U.S. law even copyrighted materials may be republished under certain circumstances. This type of usage has been vital to the level of public debate we take for granted, and must be taken into account when considering how strictly one expects to protect published material.
It is important to realize that some of the issues concerning inappropriate copying of images are not unique to the digital world. As brother Goldberg kindly reminds us in the e-mail above, any good quality reproduction of an image in a catalog or book has the potential to be “stolen”. Although this paper only describes approaches for digital images, most of the approaches also can be applied to offset reproductions.
Steps toward protection
Without protection, an organization exposes itself to litigation on behalf of the authorized agent, risks trampling the intellectual property rights of the original artist, and jeopardizes the expense of time and money that such a project requires. Opportunities for dealing with image protection surface most prominently at three key points: acquisition, presentation, and contractual use.
1. Acquisition. The moment images are first gathered is usually the best time to establish authorized use. The artist and any other relevant parties that may have claim or consideration in the use of the image should sign a release form before the image is documented. In practice, many organizations may wish to use material for which no release (or an inadequate release) exists; in those cases securing authorization from the artist or his/her estate is the proper course of action.
2. Presentation. Images must be publicly displayed in a way that both properly credits the originating artist and also clearly describes legal limitations to image re-use by viewers. In addition, technical and/or mechanical safeguards should be utilized that will limit unauthorized use. These safeguards include image resolution, watermarking, and encryption protections.
I. Image resolution
A common approach is simply to rely on displaying tiny thumbnails. These small (~1x1.5" at 72 dpi) images are adequate for getting a general impression of the item but are too low resolution for any reasonable secondary use. Of course, the degree to which such images are re-usable is subject to interpretation, creativity, and hard work. An inexpensive inkjet color printer is capable of producing remarkably good-looking copies from 72 dpi images at 100% scale. However, the file size needed for reproducing the same image at same scale for offset printing – say, for use in a brochure or book – would require a resolution of at least 300 dpi, a constraint that would considerably reduce the final size of the printed image. It is also conceivable that a simple image could be enlarged and cleaned up or re-created with applications such as Photoshop or Illustrator; however, this is usually a lot of work, presumably making it more trouble than it is worth.
Watermarked images offer a higher degree of protection for images that are of sufficient resolution to be potentially abused. They can be applied to static web images as well as material released for review or reproduction. Various commercial systems are available that allow some form of encoded identifier to permanently accompany an image and verify its source. Watermarks can either be visible or invisible.
A visible watermark (also known as “scarring”) is commonly used in applications such as photograph catalogs, allowing the viewer to see what the image is like before ordering a good copy. These embed a semi-ghosted visual logo across the image, making commercial use impractical. The advantage of this is that the mark is unmistakably bound to the viewed image and the mark itself renders the viewed image virtually useless for reproduction. One low-cost source for this is H20marker , which allows for an image or text to be overlaid onto a main image as a visible watermark at variable levels of transparency and can accommodate batch processing.
Invisible watermarks are more common in public information settings such as digital libraries and galleries. Digimarc  is one of the more popular applications; it is easy to use because its code-embedding software is included with the ubiquitous image-processing application Adobe Photoshop, and it offers an optional web-spider searching feature that helps the owner learn of new web postings that contain images with the owner's user identification. The encryption involved in embedding the unique code in the image is robust enough to survive even after resizing, cropping, and other graphic torture. Scanning an encoded image even after it has been printed out still reveals that the code persists . One limitation of this form of watermark is that it will only work with JPEG images; GIFs cannot be so marked, though this rarely affects important images.
It is important to note that watermarks cannot prevent the theft of an image, but they can be extremely valuable for tracking abuse and proving ownership once it has happened.
A more robust and elegant approach involves the use of server-side web applications. These commercial services will block unauthorized downloading of images by web spiders, may make screen capture impossible, can encrypt cached images in the browser, and prevent direct linking of images from another site. Secure Image v.2.1 by Artistscope  offers software that offers a protection firewall. Instead of the standard practice of placing an image onto the html page, the image is first encrypted and then inserted into a viewer that sits on the page replacing the image. The image converter enables a preview of the images in a window for selection, allows manipulation of size, and saves the newly encoded images or batch. Homebrew versions of this sort of protection can be built into any dynamically-generated website, such as one using a Cold Fusion server to process requests from an Access database.
3. Contractual use.
Finally, any ultimate use of an image can be carefully described by contract. Generally, use contracts include specific language regarding the following subjects, if applicable:
1. Print reproduction conditions including print run, image size, color or black and white, placement, length of use, and geographic distribution, sale/subscription price.
2. Web reproduction conditions, including restrictions on allowing direct links to the image from other sites and duration of use.
3. Image alteration, which can involve some restrictions (especially for fine artwork) that limit the end user's latitude in manipulating the image composition in any way, including flopping, cropping, and changing proportions.
4. Crediting; often there are specific guidelines as to the expected crediting of the artist, including name, date of creation, and title, as well as where that credit should reside and in what format. It is also not uncommon for supplementary information to be required, such as credit for documentary photographer, location of original artwork, image provider, participating artists, and original client.
5. Copyright protection is often specified, indicating exactly what copyright data must be displayed and where.
6. Political restrictions; some vendors go so far as to try and limit uses incompatible with their corporate mission. For example one GettyOne image of Los Angeles graffiti includes the stated restriction "Cannot be used worldwide for political causes or issues" .
One approach taken by a major institution (the Library of Congress) is to place responsibility on the patron for any misuse. This is their disclaimer:
“Division Responsibility: As a publicly supported institution the Library generally does not own rights to material in its collections. Therefore, it does not charge permission fees for use of such material and cannot give or deny permission to publish or otherwise distribute material in its collections. It is the patron's obligation to determine and satisfy copyright or other use restrictions (such as donor restrictions, privacy rights, publicity rights, licensing and trademarks) when publishing or otherwise distributing materials found in the Division's collections.
It is the Division's responsibility to inform patrons of any donor restrictions, that is, restrictions on use stipulated by the donor when the collection was transferred to the Library. These agreements between the Library and the donor about use of the collection are separate from the legal rights of copyright, publicity, and privacy discussed in this document. For example, the terms of the gift of the Work of Charles and Ray Eames stipulates that works in the collection may not be published or used commercially without the permission of the Eames Office. This requirement is not amatter of copyright--it is possible that works in the collection may be subject to copyrights held by others--but is the result of the agreement made between the donor and the Library.
In addition, the Division's staff will attempt to inform patrons about other restrictions when information is readily available ."
Candidates for Image Protection
Under which circumstances should image protection be implemented? I would suggest that any display of visual artwork that is not known to be in the public domain should have several levels of security. Even sites with materials that offer unique documentation of historic artifacts should consider basic measures. Examples of websites with where protections are weak or absent include the Free Speech Movement Archives (http://www.fsm-a.org/gallery-harawitz.html), the Pictorial Highlights from UC Berkeley Archival Collections (http://sunsite.berkeley.edu/FindingAids/highlights.html), and the Bancroft collection of Catalonian manuscripts (http://sunsite.berkeley.edu/cgi-bin/ebind2html/catalan/boethius?seq=1;cap).
The FSM site states that they are copyrighted by Howard Harawitz on the same page as the displayed images. The Pictorial Highlights page has a “Copyright © 1997 UC Regents. All rights reserved” tag at the bottom, which is linked to the Digital Library SunSITE Copyright Statement “Unless otherwise specified, copyright is held by the University of California Regents. Copying is permitted for noncommercial use by computerized bulletin board/conference systems, individual scholars, and libraries. Express written permission is required for any commercial use of the contents of this server.” The Catalonian manuscripts page contains no reference to permissions guidelines. None of these pages utilize any form of watermarking or encryption.
Images on the Web can be protected from abuse. The free exchange of ideas will only blossom as long as predators are kept at bay and information providers do not feel exploited. The many technical protection options available make it easy for information managers of both large and small institutions to assure high standards of service and control, and such measures must be included in all digitization efforts.
Lincoln Cushing, Docs Populi http://www.docspopuli.org
Special thanks to Howard Besser for his comments on the original version of this paper.
 From “Digitizing Historical Pictorial Collections for the Internet,” by Stephen E. Ostrow, Council on Library and Information Resources, February 1998 http://www.clir.org/pubs/abstract/pub71.html
 From author’s personal collection of “hate e-mail”.
 An interesting and perhaps counterintuitive consequence of this aspect of digital surrogates was raised at a presentation in 2000 by Alice Prochaska, director of digitization efforts by the British Library, in which she commented that they have had the experience of well-done digital surrogates fueling the demand for access to original artifacts.
 This includes a software distribution approach spearheaded by the GNU project, http://www.gnu.org/copyleft/copyleft.html, as well as a broader concept adopted by several publishers - see http://library.thinkquest.org/29500/copyleft.shtml
 The persistence does have limitations, but efforts by the author to squeeze the watermark out were successful only after the image was degraded to the point that its reproduction would not result in commercial duplication quality.
 see example at http://www.docspopuli.org/CubaWebCat/detail.np/detail-01.html
 see example above.
 http://www.gettyone.com/en-us/search/resultsmain.asp, (enter “eb1307-001” Search Options); “Graffiti artist at graffiti yard, Los Angeles, California, USA”, Stone photo agency, Robert Yager photographer.