Skip to main content
When you have a large catalog of products, manually adding rich detail to each product listing isn’t scalable. Image Data Enrichment is an AI-driven feature that analyzes product images alongside text to augment each product listing with highly accurate and relevant attributes. This makes every item more discoverable so you can increase conversion rates and revenue while decreasing search abandonment and lost opportunities. By automatically and intelligently adding detail to your B2B or B2C product catalogs, you can surface more results for each query, including relevant products whose original listings don’t always match how customers search.

Before Image Data Enrichment

Before data enrichment: fewer, less relevant results

After Image Data Enrichment

After data enrichment: more (and more relevant) results
With Image Data Enrichment, customers get more results for each query because richer text fields match more queries.

Image Data Enrichment with Lucidworks AI

Learn how Image Data Enrichment uses Lucidworks AI to analyze product photos and existing text, and then add keywords, synonyms, and subcategories that help shoppers find products.

How it works

Image Data Enrichment combines computer vision and natural language processing to transform sparse product information into comprehensive, searchable content. It analyzes product images alongside existing text fields to generate key words and phrases that improve findability and help customers discover products through search. As a result of this AI-driven analysis, Image Data Enrichment does the following:
  • Expands keyword coverage
  • Creates relevant synonyms
  • Adds precise product categories
The result is a complete product profile that improves discoverability across B2B and B2C commerce platforms. These are the ways to enable the feature:
  • With Fusion’s LWAI Prediction index stage
  • With the Prediction API
  • With the Async Prediction API
See Enabling Image Data Enrichment below for instructions.

When to use Image Data Enrichment

Image Data Enrichment provides measurable value when:
  • Your catalog is image-rich
  • Your catalog lacks rich text descriptions, keywords, categories, or synonyms
  • Search recall is low despite good relevance tuning
  • Manual tagging is expensive or infeasible
Image Data Enrichment is not recommended in any of these cases:
  • Images are sparse or non-existent
  • Your use case is knowledge management or site search
  • Text field are already rich
  • Compliance requirements prohibit AI-generated attributes

Examples

The following examples demonstrate how Image Data Enrichment processes real-world product catalogs and the specific enhancements it delivers.
Futsal is a popular indoor soccer variant. Imagine a customer searches your site for “futsal shoes”.The original catalog listing on the left does not match. On the right, Image Data Enrichment has added details based on visual and contextual clues that suggest this shoe is designed for futsal.

Original result

Indoor soccer shoe product photo

Toque Rebound Pro

A blue indoor soccer shoe with side lacing and a gum outsole designed for indoor courts.Keywords: blue, indoor, soccer, side lacing, gum outsoleCategories: indoor soccer footwear, soccer shoes, athletic indoor court shoesSynonyms:
  • indoor soccer shoe indoor court shoe
  • traction grip

Image Data Enrichment result

Indoor soccer shoe product photo

Toque Rebound Pro

A blue indoor soccer shoe with side lacing and a gum outsole designed for indoor courts.Keywords: blue upper, woven upper, side lacing, clean striking surface, gold embroidery, knit collar, gum rubber outsole, herringbone tread, soccer, futsal, indoor, indoor playSynonyms:
  • indoor soccer shoe indoor court shoe
  • traction grip
  • woven upper textured upper
  • gum rubber outsole gum sole
  • ball control touch control
  • stability support
Categories: indoor soccer footwear, soccer shoes, athletic indoor court shoes, futsal shoes, indoor soccer footwear, futsal performance shoes, court traction footwear, technical soccer shoes, low profile indoor trainers, gum sole indoor shoes

Benefits

For your team

  • Higher conversions and revenue
  • Faster adaptation to the latest search trends
  • Improved performance against key business metrics

For your customers

  • More (and more relevant) results for every search
  • Easier product discoverability
  • Higher satisfaction

Enabling Image Data Enrichment

You can enable this feature in Fusion or by using the Prediction API or the Async Prediction API. Before you get started, ensure your product catalog images are publicly available with a URL. Alternatively, you can provide base64-encoded images.

Using Fusion

In Fusion, you enable Image Data Enrichment by configuring the LWAI Prediction stage in your index pipeline. You might also need to configure the Field Mapping index pipeline stage and the Query Fields query pipeline stage to generate and display metadata for images. It is essential to review your Image Data Enrichment results. Plan to test different combinations of “images only” or “images with text fields” to find the balance that produces the best results.
  1. In Fusion, open the index pipeline you want to use for Image Data Enrichment.
  2. Click Add a new pipeline stage and select LWAI Prediction.
  3. In the Label field, enter a unique identifier for this stage. For example, LWAI Image Metadata Enrichment.
  4. In the Condition field, enter what the document must contain for the stage to select it. For example, doc.hasField("image_url_t") means the document must have the image’s web address stored as a text field to be included in the stage.
  5. In the Account Name field, select the Lucidworks AI API account name defined in Lucidworks AI Gateway.
  6. In the Use Case field, select or enter image-metadata-enrichment.
  7. In the Model field, you must enter a multi-modal model, which is required by the image-metadata-enrichment use case. A multi-modal model can read images as well as text.
  8. In the Input context variable variable field, enter the name of the field that contains the image URL. For example, <doc.image_url_t>.
  9. In the Destination field name and context output field, enter the name that will be used as both the field name in the document where the prediction is written and the context variable that contains the prediction. For example, image_enrichment.
  10. In the Use Case Configuration section, you can add useCaseConfig parameters and values to send to Lucidworks AI. These parameters give the model context, such as a title and categories, and set the maximum number of keywords, synonyms, and subcategories it returns. They are not filters: the stage enriches every document that meets its condition. To control which documents the stage processes, use the Condition field. The stage sends each value exactly as you enter it, with every request. It doesn’t replace <doc.field> templates in this section with values from the document, so enter fixed values that apply to every document the stage processes. Enter categories as a JSON array, such as ["bathroom","kitchen","home","industrial"]. For the parameters and their defaults, see image metadata enrichment parameters. The maxSynonyms limit applies to each keyword, so the defaults of 15 keywords and 5 synonyms per keyword can return up to 75 synonyms for each document. Lower these limits if the generated terms are too broad. This example sends a title and a list of categories as context for the model, and sets locale to en-US so the generated terms are in US English. It also limits the output to three keywords, five synonyms per keyword, and three subcategories. Image metadata enrichment use case configuration In this screenshot, the title value is a <doc.ProductDescriptionDetail_t> template, which the stage sends as literal text. Replace it with a fixed value.
  11. In the Model Configuration section, you can add model parameters to send to Lucidworks AI, such as temperature or maxTokens. For Google Agent Platform models that use your own API key, googleRegion and googleProjectId set the Google Cloud region and project where the model runs.
  12. You can run this stage in two ways:
  • Do not select the Call Asynchronously? checkbox to process this stage completely before proceeding with indexing. This is the same as using the Prediction API.
  • Select the Call Asynchronously? checkbox to process predictions asynchronously without blocking the pipeline. This is the same thing as using the Async Prediction API, and is useful to process large documents without impacting further indexing processes.
  1. In the API Key field, enter the secret value specified in the external model.
  2. Click Save.
You may also need to set some parameters in the Field Mapping index pipeline stage.
  1. Click the Field Mapping stage and scroll to the Field Translations section. For example, if the source field names are extremely long, you can use the Field Translations section to shorten them for your target index fields. Image metadata enrichment field mapping stage
  2. When you enter all of the parameter and field values, click Save.
Run a datasource job that uses the configured index pipeline to add enriched data to your index. When the datasource job finishes, you can see the image metadata data fields in the Query Workbench.
  1. In your query pipeline, select the Query Fields stage.
  2. Scroll to the Return Fields section and enter the image metadata fields to include in the query results. The fields need to include, but do not have to be limited to, the fields specified in the indexing stage. This example includes the fields specified in indexing field example. Query Workbench Return Fields
Review the results in the Query Workbench, which should include:
  • The image link under the title or the image URL to view the image.
  • The title, the locale of the generated terms (en-US in this example), and the calculated score of the item.
  • The keywords, subcategories, and synonyms generated for the image. For this example, the categories sent were bathroom, kitchen, home, and industrial. The model uses these categories as context for the subcategories it generates. They don’t filter which documents are enriched. Image metadata enrichment query results

Using the Prediction API

If you’re not using Fusion, you can enable Image Data Enrichment by making calls to the Prediction API and incorporating the response into your data source.

Example requests

The Prediction API requests below use image metadata enrichment parameters.

Example response

The response for all of the above examples looks like this: