Audience Measurement Engine
Try before you buy:
Read code documentation:
Brief description
The Audience Measurement Engine is a ready-to-use C++ / WebAssembly computer vision solution that measures how many people look at a product in a shop window or in-store display, how long it holds their attention, and what emotional reactions it triggers, aggregated across six emotion categories.
Detailed description
This engine is designed for brands, retailers, and digital signage providers who want to understand how real people react to products and creative content in front of a camera. Using face detection, gaze estimation, and facial emotion recognition, the engine turns raw video frames into actionable engagement metrics that can be fed into your dashboards or analytics pipeline.
For each viewer detected in a configured region of interest (for example a specific product in a shop window or a promotional stand), the engine continuously estimates whether the person is actually looking at the target and for how long. Over time, it builds up statistics that answer three key questions:
- How many people looked at the product? – viewer counts and dwell time in front of the display.
- How strongly did the product capture their attention? – attention estimation based on head and eye gaze direction, distinguishing true viewers from passers-by.
- What emotional response did it generate? – total viewing time accumulated in six pre-trained facial emotion categories.
Under the hood, the Audience Measurement Engine builds on OpenCV and neural-network based components for face detection, dense facial landmarks, gaze estimation and emotion classification. Results are aggregated over time so you can analyse engagement per session, per campaign, or per product placement.
The project ships with full C++ source code and can be compiled to WebAssembly using Emscripten, allowing you to run the same engine directly in the browser. This makes it suitable both for back-end batch analytics (processing recorded video) and interactive browser demos where customers or stakeholders can experience the system live.
Whether you want to optimise a storefront, compare A/B variants of creative content, or build your own analytics product on top, this engine gives you a solid, well-documented foundation that you can extend and integrate into your existing systems.
What's Included:
- Complete C++ source code files
- Detailed code documentation
- Pre-trained neural networks
- Precompiled OpenCV WebAssembly Module
- Instructions for compiling the project with Emscripten
- Bonus: Build guide for OpenCV that shows how to compile it to WebAssembly
Technical Requirements:
- Knowledge of C++ programming and basic understanding of neural networks.
- Emscripten SDK for compiling to WebAssembly.
Why Invest in My Solutions?
Investing in Computer Vision solutions from my website is a smart choice if you’re looking to integrate technology into your business efficiently and affordably. Developing Computer Vision solutions from scratch is both time-consuming and costly. My website offers pre-built solutions that you can start building with immediately, saving you both time and money.
When you purchase from my website, you receive the complete source code, comprehensive documentation, and all necessary assets, giving you full control and flexibility to customize the product to your specific needs. Additionally, you benefit from lifetime access and the freedom to use the solutions commercially, allowing you to adapt or sell your own products built with them.
In short, purchasing from my website provides you with a cost-effective, time-saving solution and extensive documentation, making it a practical and profitable investment for your business.
License
By purchasing or downloading any project, you agree to the full license terms, which you can view here: LINK