Case study
Sneaker Visual Retrieval System
- Client: AlwaysLegit
- Industry: Retail & E-commerce
- AI capability: Computer Vision
- SKUs searchable from a single photo
- 150K+
- Retrieval time per query
- <1s
The challenge
150,000 SKUs, one photo, and an expo floor in the background
The primary challenge was the scale and speed required. Matching a single photo against 150,000 products in under a second is computationally intensive. Furthermore, photos taken at busy expos often have cluttered backgrounds, varying lighting, and multiple angles, making traditional image matching unreliable. The system needed to be robust enough to handle these real-world variations while maintaining extreme accuracy to ensure the correct SKU was identified.
What we built
Visual search that answers in under a second
Find the shoe, turn it into an embedding, search the catalogue.
- Stage 1
Isolate the sneaker in the photo
A sneaker-specific detection model finds the shoe and crops away the expo background: people, booths and surrounding products.Clean sneaker crop - Stage 2
Turn the crop into an embedding
Deep learning feature extraction produces a high-dimensional embedding of the shoe's structure and colour distribution: panel layout, shape, branding patterns and colour signature.Shape and colour embedding - Stage 3
Search the SKU catalogue
The embedding is matched against a vector database holding representations of more than 150,000 sneaker SKUs, using approximate nearest neighbour search built for that scale.Ranked candidate SKUsVector databaseApproximate nearest neighbour - Stage 4
Return the product record
The closest match comes back with its product details, so the team on the stand can answer in seconds instead of searching a catalogue by hand.Matched SKU in under a second
The visual search dashboard. For this query, detection took 0.04 s and the database search 0.58 s.
The impact
Anyone on the stand can identify a shoe from a photo
- 150K+ SKUs searchable from one photo
- Under a second per query
- Works on cluttered expo photos
The activation transformed how AlwaysLegit interacted with enthusiasts at events. Attendees could take a photo of any sneaker, and the system would instantly provide the exact product details, pricing, and availability. This not only provided a 'wow' factor but also significantly improved the efficiency of their on-site team, leading to higher engagement and a larger volume of SKU lookups than previously possible with manual search.
Keep exploring
Related case studies
Real Estate & PropTechComputer VisionAI Property Image Intelligence for Listing Optimization & Compliance
PropTech founders and real estate leaders understand that property photos are the most critical asset in closing deals. PropTexx needed an automated way to enforce strict listing compliance and extract valuable insights from millions of unstructured real estate images without slow manual review.
- Property features tagged automatically
- 50+
- Compliance violation types detected
- 15+
Sports & MediaComputer VisionSoccer Match Analytics & Insights
The aim of this project is to give soccer analytics of matches being played. The goal is to build an AI tool that takes as input a video of a soccer match and then analyzes each player and extracts different analytics. This included: 1. Player and ball detection, 2. Player and ball tracking, 3. Player re-identification, 4. Field detection, 5. Event detection including passes, intercepts, goals etc.
Multi-stage CV pipeline. Real-time event detection. Player re-identification.
Read the case study
Health & FitnessComputer VisionX-ray to 3D CT Reconstruction & Knee Alignment Analytics
Traditional CT scans are several times more expensive than X-rays, often delaying critical orthopedic diagnosis. Our client needed a system to reconstruct 3D CT-grade results from standard 2D X-rays to calculate precise bone alignment angles for surgical planning.
2D to 3D reconstruction. Automated angle calculation (CPAK, mHKAA). Enhanced with GANs.
Read the case study
Book a strategy session
Talk to an AI engineer about your project
Tell us what you want to automate. The first call is a 30-minute working session with an engineer, not a sales pitch.
- Send the form, it takes 2 minutes
- We reply within 1 business day, under NDA if you need it
- A 30-minute call to scope feasibility and next steps