Brandshight AI

BrandSight is an AI tool that checks store visuals against brand guidelines, detects errors in logos, colors, or placement, and gives scores & insights to keep every display 100% on-brand.

Comment

BrandSight AI – Visual Brand Compliance Engine

Inspiration

Global brands invest millions in perfect visuals — logos, colors, store layouts — but once campaigns go live, they lose control over how their branding appears in the real world.
Manual audits are slow, error-prone, and expensive. We wanted to create a system that automates visual compliance checks and provides actionable insights to improve brand consistency.

What it does

BrandSight AI compares brand reference images with real-world store photos to:

  • Detect misplaced or missing products
  • Identify wrong logos, colors, or outdated signage
  • Generate a compliance score for each display
  • Highlight non-compliant areas with heatmaps
  • Provide actionable insights to improve store layouts
  • Track compliance trends across stores and campaigns

How we built it

  • Frontend: React.js + Tailwind for a clean dashboard and upload interface
  • Backend: FastAPI (Python) for handling image processing and analytics
  • Visual Difference Engine: OpenCV + SSIM to compute pixel-level differences
  • AI Classification: YOLOv8 / pretrained CNN for detecting logos and display issues
  • Storage & Hosting: Firebase / MongoDB Atlas for images and data, hosted on Vercel + Railway

Challenges we ran into

  • Aligning reference and store images despite different angles and lighting
  • Balancing sensitivity of visual difference detection to avoid false positives
  • Displaying clear heatmaps that are intuitive for non-technical users
  • Incorporating actionable AI insights in a simple, understandable format

Accomplishments that we're proud of

  • Built a functional AI-powered visual compliance engine in just 10 days
  • Generated real-time visual diff heatmaps with compliance scores
  • Added preliminary AI suggestions for improving brand displays
  • Designed a dashboard that allows multi-store compliance tracking

What we learned

  • Computer vision techniques like SSIM and YOLO can be combined to detect real-world compliance issues effectively
  • Presenting AI results visually (heatmaps, scores) greatly improves usability
  • Hackathon projects benefit from scoped MVPs that showcase the core functionality clearly
  • Combining rule-based insights with AI predictions creates actionable recommendations without heavy training

What's next for BrandSight AI

  • Multi-store batch analysis with automated reports
  • Mobile app for real-time image capture and compliance checks
  • Predictive AI suggestions for improving future campaigns
  • Generative AI previews to visualize ideal in-store layouts
  • Scaling to multiple brands and campaigns for enterprise use

Built With

Share this project:

Updates