MIRKO

Image Reco — Shelf Recognition with Computer Vision

A data-driven solution that uses computer vision to analyze store shelves, providing actionable insights for trade promotion and category management for tobacco products.

Industry
Retail
Services
AI & Data, Computer Vision, Mobile Development
Technologies
Computer Vision, TensorFlow, Python, Salesforce
Image Reco — Shelf Recognition with Computer Vision

Overview

About The Project

Image Reco is a tailored, data-driven solution that uses cutting-edge computer vision technology to analyze store shelves, provide actionable insights and precise data to support trade promotion and category management, boost profitability and build customer loyalty.

Image recognition is embedded into field sales and retail execution applications, allowing users to scan shelves and identify incorrect facings, out-of-stocks or missing items in distribution.

The Challenge

Audits That TookToo Much Time

Our mission was to create an innovative merchandising tool that helps drive incremental sales and increases productivity, improves shelf condition insights, saves up to 90% of time spent on in-house audits and decreases human resources expenses.

Audits That Took
Too Much Time

Our Solution

Tobacco SKU RecognitionIntegrated with Salesforce

We designed a 100% customized solution using computer vision for tobacco product recognition, with 85% recognition accuracy on decent photos, 2–4 second photo processing and the ability to track 200+ SKUs.

Fully integrated with Salesforce, Image Reco lets the owner check on-shelf availability in real time, control planogram compliance and receive alerts to take corrective action quickly.

Tobacco SKU Recognition
Integrated with Salesforce

Key Features

What WeDelivered

The Impact

Results ThatMatter

Faster, cheaper and more accurate in-store audits.

85%

Recognition accuracy

2–4 s

Photo processing

200+

SKUs tracked

-90%

Audit time