RollingDot MegaGrants Application

Unreal Engine’s point cloud plugin evaluation

This first video illustrates the rendering obtained when evaluating the UE’s point cloud plugin. We compare these results with those obtained using our new prototype including the first version of our data structure. As we can see the resolution is better handled and we are able to avoid its visible artefacts. The « Capitole » model is 100 188 944 points while the « Zenith » model is 94 500 016 points.

RollingLive with Unreal Engine

This second video showcases our RollingLive prototype powered by UE5 to visualise and process in real time massive 3D point cloud data. Our unique Collaborative AI capabilities allow users to optimize 3D data processing (classification) without having to access the code.

As shown in the video, users can visualize and interact with large acquired point clouds in a better way than with the UE’s Lidar plugin.

In addition, users can annotate different classes in order to train the AI model. Little data is needed to train our agile & fast AI models. These also process the full datasets in record times in order to present the prediction results within the interface. 

Thanks to the use of UE, we allow users to interactively improve our fast AI models within the real time viewer without having to access our machine learning codes.