The science behind NeuroVision
About
NeuroVision is built on three core components:
- A vast amount of eye-tracking data that Neurons has accumulated over the years. With no less than 12,000 participants tested and over 500 million data points, this database is among the world’s largest on consumers’ visual attention.
- The latest Machine Learning approaches, through TensorFlow, where multiple algorithms have been tested in parallel to produce the most predictive models.
- The use of computational neuroscience to produce a visual saliency model that shows the most visually salient features in an image.
High value low cost
predicting eyetracking
| EYE-TRACKING | NEUROVISION | |
|---|---|---|
| COST | High | Low! |
| SPEED | Slow | Fast! |
| ITERATIONS | Few | Unlimited! |
| COMPLEXITY | Difficult | Easy! |
| ACCURACY | High | High! |
Vision
Making the attention AI
NeuroVision is based on two main ingredients: one of the largest eye-tracking databases, and state of the art machine learning models.
Eye-tracking database
With a database based on well over 12.000 participants, Neurons holds one of the largest single databases of high quality eye-tracking data. This is a critical element in training machine learning models.
machine learning
The NeuroVIsion Attention AI solution has trained 30 different machine learning models, each models running for weeks to months. The winning model predicts eye-tracking with over 90% accurac
High value low cost
predicting with 90% accuracy
