Automatically select the most engaging content

Neon Open is an open-sourced artificial intelligence platform that predicts which images and videos will drive more clicks, likes, and shares. ​

use cases

Computers can see. With Neon Open they can feel. 

Neon Open is the world’s largest and most comprehensive deep learning model and dataset that quantifies the perceived emotion of images.
​It can be used in conjunction with state-of-the-art object detection models to extend the capabilities of visual AI from object detection and scene recognition to predict a real human emotional response. 

OBJECT DETECTION + HUMAN EMOTION == FUTURE OF VISUAL AI

Add emotional intelligence to your AI

At Neon Open we not only share our models and source code, we share the images and raw training data used to train our deep neural networks. We outline the experimental tasks used to collect training data so that engineers, academics, artists, and individuals can expand the training set.

Dataset

5 million Images tagged with emotion data collected through crowd-sourced experiments

Model

Fully trained CDNN running as a modified version of Inception V3 on TensorFlow

Model Serving Code

Source code that allows you to deploy the Neon Open model in a production environment 

Check out our model and code on GitHub

FORECAST VISUAL TRENDS

Learn image style

Neon Open goes beyond standard object detection to learn the style of an image. Our deep learning models can be used to predict an individual's style or preference across different objects from water glasses to watches. This same capability can be used to identify images with a consistent visual identity for brands.

Learn more

Grow revenue with more engaging images

"Not only did Neon take a significant amount of work off our hands, it also improved the 
clickability of our thumbnails by 30%"

- JIM HALL, IGN ENTERTAINMENT​

See how

Together lets make the future open

At Neon Open we believe that for artificial intelligence to improve the lives of individuals there needs to be an open sharing of data and a free transfer of ideas.

​By making our data, models, and methods freely available we welcome journalists, artists, engineers, academics, creatives, and non-profits using photos, video, and mixed reality to learn more about the data behind artificial intelligence. 

Learn more

Join the conversation at #neonopen