Fashion Tagging

Fashion Tagging is a visual AI service that automatically recognizes fashion products in images, providing their category, subcategory, and various tags. An optional meta endpoint tags the background, scene, view, and body parts of the person wearing the items, which helps with the selection of images for the product listings and galleries.

Your fashion image is processed by a group of object detection and image recognition models working together. First, all fashion apparel and accessories are detected. Then, each item is categorized and tagged accordingly (e.g., shoes are tagged for sole, heel, material, and type). The main color of each recognized object is also extracted by default, eliminating the need for a separate color extraction model unless detailed color analysis is required.

Most fashion e-shops and comparators use Fashion Search, which combines Fashion Tagging with Object Detection and Product Recommendations (visual search). To use Fashion Search, your collection is synchronized to Ximilar cloud, where each picture is analyzed only once and then discarded. You will get categories, tags, colors, and similar items to each of your images under one API request.

The fashion taxonomy of Ximilar is public in several places. You can:

• Get a PDF file with full Ximilar Fashion Taxonomy via the Fashion Tagging page.
• Try out the public interactive demo and see how it works on your images.
• Browse the API documentation.

Automated Fashion Tagging is used on fashion product images of e-shops, price comparators, fashion brands, and specialized collections. It is based on numerous image recognition tasks trained to recognize separate product categories and object detection models. That is why it works on both single product images and more complex images, including user-generated content or social media images.

Yes! Ximilar has a free and public Fashion Tagging demo. You can either upload images or their URLs and see for yourself how automatic fashion tagging works.

You can also use Fashion Tagging in our App. See our Pricing for details. If you have large volumes of images to be processed every month or need customization, contact us to discuss a custom plan.

Fashion Tagging labels your fashion items, assigning categories (e.g., skirts), subcategories (e.g., A-line skirts) and tags (for color, design, pattern, length, rise, style…). By default, it provides data for one main object in an image. Endpoint meta can also provide tags for the photography background, scene, or body part in the fashion image.

Fashion Search is an all-encompassing solution, wrapping all typical fashion AI services into one. It integrates:

• Fashion tagging, including Dominant colors
• Object detection for precise labelling and search of individual items
• Visual Search, recommending similar items from your collection

Both Fashion Tagging and Fashion Search include color analysis. The colors are supplied as tags and can serve for filtering and search on your website.

I only need a single fashion AI solution

All our fashion AI solutions can also be employed individually. Examples include product similarity, search by photo (reverse image search), fashion apparel detection, or color-based search.

Fashion Tagging can be combined with object detection to categorize and tag individual items in a more complex fashion image. That is why our solution Fashion Search automatically detects apparel, footwear and accessories in your images, provides tags, and finds the most similar products or images.

These fashion services work on both product images and real-life photos, e.g., fashion influencer pictures. Endpoint meta can also optionally provide tags for background, scene, view, and body part of the person wearing the items.

Fashion Tagging is one of our most complex ready-to-use services. It works with over a hundred recognition tasks, hundreds of labels, and dozens of fashion attributes.

It identifies the top category of product (e.g., accessories, bags, jewellery, watch, clothing, underwear, footwear), then the category (e.g., accessories/belts), and its features such as colour, design, material, or pattern.

Customization

There is also an optional meta endpoint for tagging of background, scene, view, and body part of the person wearing the items.

If you miss any important attributes, the taxonomy can be adjusted to fit your use case. It can also be used in different languages than English.

Level up with Fashion Search

Furthermore, our service Fashion Search combines deep tagging with object detection to ensure all fashion items in an image are tagged. The detected objects are then automatically used for similarity search in your collection.

With Ximilar, you can customize the tagging taxonomy:

• Custom taxonomies can be applied through taxonomy mapping.
• The tagging services can also be easily switched to different languages.
• We can also replace or add tags on demand.

The first steps could be:

• trying how the service works in our public demo or App
• downloading the full Ximilar Fashion Tagging taxonomy
• checking the API documentation for details

If you do not find the attributes you need, contact us to modify the service to fit your use case.

You can also use our Computer Vision platform to train your own custom categorization and tagging models and combine them with ready-to-use solutions.

We offer multiple options for dominant color extraction that you can select from. The outcome is provided in a structured format, usually JSON.

Dominant product vs. whole image

The product endpoint allows you to extract colors from a single dominant object in an image (product photo), whereas the generic endpoint extracts the dominant colors from the entire image, a mode typically used in stock photography.

Basic color for searching & filtering

This mode identifies one main color of the dominant object out of a total of 16 basic colors. The extracted color can be utilized as an attribute for filtering and searching fashion items.

Pantone palette: detailed color analysis

This mode provides a group of dominant colors, their hex codes, the closest Pantone name, and coverage of the image in %. It is ideal for similarity search (search by color).

Yes, you can! Automated Fashion Tagging works on product images as well as real-life photos. Our tagging combines object detection, which identifies fashion items in an image, with image recognition, which categorizes these items and provides you with tags.

The generation of product titles and descriptions can be easily automated with Ximilar AI.

Tags extracted from Fashion Tagging can indeed be used. These tags can serve as input data, assisting the AI in understanding and generating relevant and accurate product titles and descriptions.

You can select the tonality and the length of the texts, and also set some basic rules for their generation. Contact us via live chat or contact form for details.