Backlogs of cataloging are a common challenge for many museums. These backlogs are often time-consuming, and because they are less urgent than loans or exhibitions, they tend to be sidelined in favor of other priorities. Addressing these backlogs requires planning, especially when a museum depends on a small staff and a team of volunteers. Implementing the right strategy is crucial for maintaining efficiency.
In small museums, overworked staff often supervise volunteers, making it essential to streamline the cataloging process. Simplifying the workflow by starting with just an image, title, and description is an effective way to launch any cataloging project.
CatalogIt has introduced the ‘Describe Image Assistant’— an AI-powered tool designed to provide users with an efficient and consistent solution for automating image descriptions. Since volunteers may lack the expertise to write detailed text for every image in the museum's collection, this tool can generate descriptions quickly, freeing up the museum’s limited resources to focus on other essential tasks.
Using AI for Visual Descriptions
The Describe Image Assistant uses AI to generate descriptions for any image in your CatalogIt account and can be useful for creating general-purpose descriptions of images, saving you time and improving efficiency, providing a streamlined way to create content for records. This text can be used in object labels, social media posts, and as a tool to help efficiently catalog lots of object records in a short period of time.
These descriptions can be used as visual descriptions or alt (alternate) text, a written description of an image to make visual content accessible to users who are visually impaired or using screen readers. For records published to the internet, the image description also aids in search engine optimization( SEO) by enabling search engines to understand and index the image content.
Using an AI-powered alt text tool can be particularly beneficial for a small museum with primarily volunteer staff for several reasons:
- Efficiency: An AI tool can generate descriptions quickly, allowing the museum staff to focus its limited resources on other tasks.
- Accessibility: An AI tool makes it easier for all images to have alt text, helping the museum meet accessibility standards and reach a broader audience.
- Consistency: Volunteers may have varying levels of experience cataloging, leading to inconsistencies in quality. An AI tool provides a standardized approach, ensuring that all descriptions are comprehensive.
- Scalability: As the museum's collection grows, the task of writing alt text manually becomes more daunting. AI can scale with the collection, generating descriptions for new items without a significant increase in workload.
Starting the Cataloging Process with your Volunteers
Assigning a group of objects to specific volunteers is a great first start. Ensuring that the same volunteer sticks with the same group of objects will help with consistency and follow-up by staff later. Once the group is assigned, supervising staff can train the volunteer(s) on the specific workflow and process. Ideally, all records should include:
- An image (or multiple in the case of a three-dimensional object)
- A title or object name
- Descriptive text
- Dimensions
- Any obvious condition issues
The visual description should include as much information as can be gleaned from looking at the object, including material, color(s), description, function, etc. It can be written manually or by using the CatalogIt Describe Image Assistant. Since this feature is AI-generated, it is important to proofread to ensure accuracy. It is also best to take this time to physically measure your object because the AI cannot extrapolate dimensions.
Having your volunteer stick to the same group of objects each week, rather than having different volunteers working on the same list, will help to ensure nothing falls through the cracks. After each group of objects is cataloged, staff should review the records and work with the volunteer to safely store these objects.
The records should be flagged as “partially cataloged” so that they can be worked on later, as time permits, to comprehensively catalog each item. These records can be cataloged without handling them again since the CatalogIt entry includes the image, visual description, and dimensions.
Using CatalogIt’s Describe Image Assistant
Using the Describe Image Assistant is simple. The image description field is just below the caption included beneath every image you upload into CatalogIt.
When editing Entries and Profiles, click the "document" icon at the right of the Image Description field:
The Describe Image Assistant tool will open:
Specify the length of the description you want to generate. If the long or short options are not providing satisfactory results, you can specify a custom length between 10 and 200 words. Optionally give the system a hint of the context and content of the image. You should only need to specify a hint if the image has ambiguous content or the generated description is not accurate.
Click the Describe button. After a few seconds, the tool will close and the image description will appear in the Image Description field. Proofread the description and make any necessary edits. You may also want to copy the text to the Entry Description field. Save the Entry or Profile.
For example, using the object pictured above, multiple descriptions were generated using different prompts. Below are the unedited AI-generated responses based on the criteria selected in the Describe Image Assistant:
Short Description:
“The image shows a collection of various knotted ropes displayed on a wooden board.”
100 Word Description:
“This image displays a collection of neatly tied knots arranged side by side on a dark wooden board framed with a lighter wood, appearing as an educational or decorative display. Each knot is distinct in design, demonstrating various techniques and purposes, secured on the board by small nails, which indicate a systematic approach likely aimed at teaching or showcasing the art of knot tying.”
200 Word Description (Custom):
“The image displays an extensive assortment of different knots, neatly arranged and mounted on a rectangular, dark wooden board. Each rope knot varies in design and complexity, showcasing a variety of shapes and structures. The board serves as both an educational display and a decorative piece, highlighting the artistic and functional aspects of rope tying. The background is neutral, focusing attention on the intricacy and craftsmanship of each tied rope. This collection could be used for educational purposes, such as teaching knot-tying skills, or simply as a unique wall decoration. The overall presentation is orderly and aesthetically pleasing, emphasizing the skill involved in creating each individual knot.”
Custom Description with Prompt:
Prompt: Japanese Fisherman Knot Board
“The image displays a collection of Japanese fisherman knots intricately tied and mounted on a dark wooden board. Various knot types are showcased, each clearly different from the others, ranging in style from looped and coiled to interwoven patterns, neatly aligned in horizontal rows to create an organized and educational display. The natural color of the ropes against the dark wood background highlights the craftsmanship and aesthetic appeal of traditional knot-tying skills.”
These responses demonstrate that the Describe Image Assistant Tool is an invaluable feature in CatalogIt, significantly boosting efficiency and enabling even the smallest museums to create consistent, more complete object records.
The image description can also be copied and pasted into any other field in CatalogIt. This is a new and evolving feature and in the future may be incorporated into other fields in CatalogIt to add to the efficiency of working in your collections database.