Wideeyes link
Feature #1 - Search By Image (Similar image search :))
Input - Image
Technical details
- 1000 different attributes, thus is able to connect different types of fashion images Real-time, fully automatically
Result - Results of the most similar in-stock matches of outfit
Feature #2 - Similar Recommendation (Transaction-based similar items)
Input - Most similar products to the one the customer is viewing
Technical Details
- Personalize the results for every shopper (size, brand style, etc.)
- Business rules (price range, campaigns, stock by country, area, store, etc.).
Result - Similar images based on current browsing data and metadata configuration
Feature #3 - STYLE ADVISOR (Similar to cross-category bundle)
Input - Browing / Purchase databased
Technical Details
- Fully automated complete outfit based on a single product.
- Perfect outfit recommendations and increase the shopping basket.
Result
- Worn together, Fit together recommendations
- Complete The Look
Feature #5 AUTO-TAGGING
Input - Image
Tech Details - 300 high-quality tags based on images, adapted to more than 60 categories (apparel, jewelry, and fashion).
Attributes - Gender / Category / Color / Shape
Result - Autotagging recommendations and Automating noise detection in image
My Feedback - Basically good amount of detection / classification / styles / category mapping. They have models built but customized based on customer dataset
Memory-efficient embeddings for recommendation systems
Keep Exploring!!!
No comments:
Post a Comment