Visual image search

In the battle for consumer attention, speed and ease are essential. Users are overwhelmed with information. Relevant search and discovery can help them sift through the noise.

Searching with images is convenient when it’s hard to find the words to describe what you’re looking for or when it’s too time-consuming to search and filter for what you had in mind. Uploading an image as a search query lets your users quickly find information or buy a product.

It’s becoming more common for search and discovery experiences to include this capability, especially on ecommerce apps and sites. Some businesses promote this capability even in brick-and-mortar stores. Users can snap pictures of items in the store and use the images to find more information in the catalog.

You can add searching by images to your existing Algolia implementation.

Image classification

Though searching by image may be your end goal, the first step is to enrich your Algolia data based on the images you already have. In other words, you need to add textual classifications to your records based on your images.

You can retrieve classifications from AI-powered image recognition platforms and add them to your Algolia records. Then when a user searches using an image (at query time), those same classifications can help retrieve the correct record.

For example, if you run a fashion ecommerce site, you can automatically extract features like the type of neckline, sleeve length, colors, and patterns from your product images.

Without first enriching your Algolia records, when a user provides an image as a search query, it won’t be able to match relevant records with the same or similar image. Even if you don’t plan on implementing search by image, image classification and tagging can offer a more relevant and discovery experience in and of itself.