In the past, 3d has been used mostly for games movies and films, and these are usually big budget items that can afford specialists working on basically, specialist tools, so we want to do is basically open up this whole area to non specialists and really make it much Faster and easier to generate 3d content, opening up several bottlenecks in different industries. Some of the enablement is in the e-commerce area and in this area you have customers who want to make their products are great online. So often in
The case in the past that means going to a photo studio or hiring a professional to do a photo shoot of your items and put them in Nice scenes. This is a slow and expensive thing. What we want people to be able to do is through a small, in-house and creative or by themselves be able to make really good-looking
content for their brands so this can be on branded consumer items and like packaging or it can be right up to and things like furniture so we want to enable people to show their things as best possible online one thing to be really careful of and that we did learn is that in the AI space if you cant automate everything then you really have to think about very carefully how to fit humans in the loop so that’s going to be a challenge across any AI industry so this is an internal tool that we use.
to build 3d models as fast as possible in reproduce the way so what that means is once we build something using a graphical programming approach we can use that over and over and this really helps us speed up the model building process so this is a very simple example of a simple cup like a glass-like shape and the power here is that this shape is fully parametrized so we can basically change in attribute over on the left and that’s reflected in a change of the.
shape of the model so this allows us basically to bring a human into the loop especially in the fully autonomous machine learning goal in the future so after our early monolithic staff we started moving out to basically having a static file storing on s3 we move to database anon or D s and we also have a cloud-based rendering and deep learning setup where we basically use a queueing system to launch GPU based P 2 instances so this system works both on our image enhancement to enhance towards AI and.
also works on our in-house cloud rendering so to find out more you can go to our website zag AI is a BGA I and basically therell be some useful demo videos on there and you can contact us through the contact button on the side.
Leave a Reply